Google Git
Sign in
apache / derby / dbed020d3949dd7707a6eca23741aa5ba155ece1 / . / java / engine / org / apache / derby / impl / sql / compile / ResultColumnList.java
blob: c64f2ff5ab9df483ce9abb798c9b87d852273263 [file] [log] [blame]
/*
Derby - Class org.apache.derby.impl.sql.compile.ResultColumnList
Licensed to the Apache Software Foundation (ASF) under one or more
contributor license agreements. See the NOTICE file distributed with
this work for additional information regarding copyright ownership.
The ASF licenses this file to you under the Apache License, Version 2.0
(the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package org.apache.derby.impl.sql.compile;
import java.lang.reflect.Modifier;
import java.sql.ResultSetMetaData;
import java.sql.Types;
import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import org.apache.derby.catalog.types.DefaultInfoImpl;
import org.apache.derby.iapi.error.StandardException;
import org.apache.derby.iapi.reference.ClassName;
import org.apache.derby.iapi.reference.SQLState;
import org.apache.derby.iapi.services.classfile.VMOpcode;
import org.apache.derby.iapi.services.compiler.LocalField;
import org.apache.derby.iapi.services.compiler.MethodBuilder;
import org.apache.derby.iapi.services.context.ContextManager;
import org.apache.derby.iapi.services.io.FormatableBitSet;
import org.apache.derby.iapi.services.loader.ClassFactory;
import org.apache.derby.shared.common.sanity.SanityManager;
import org.apache.derby.iapi.sql.ResultColumnDescriptor;
import org.apache.derby.iapi.sql.conn.LanguageConnectionContext;
import org.apache.derby.iapi.sql.dictionary.ColumnDescriptor;
import org.apache.derby.iapi.sql.dictionary.ColumnDescriptorList;
import org.apache.derby.iapi.sql.dictionary.ConglomerateDescriptor;
import org.apache.derby.iapi.sql.dictionary.DataDictionary;
import org.apache.derby.iapi.sql.dictionary.DefaultDescriptor;
import org.apache.derby.iapi.sql.dictionary.TableDescriptor;
import org.apache.derby.iapi.sql.execute.ExecPreparedStatement;
import org.apache.derby.iapi.sql.execute.ExecRow;
import org.apache.derby.iapi.sql.execute.ExecRowBuilder;
import org.apache.derby.iapi.store.access.ConglomerateController;
import org.apache.derby.iapi.store.access.StoreCostController;
import org.apache.derby.iapi.store.access.TransactionController;
import org.apache.derby.iapi.types.DataTypeDescriptor;
import org.apache.derby.iapi.types.DataValueDescriptor;
import org.apache.derby.iapi.types.RowLocation;
import org.apache.derby.iapi.types.TypeId;
import org.apache.derby.iapi.util.JBitSet;
/**
* A ResultColumnList is the target list of a SELECT, INSERT, or UPDATE.
*
* @see ResultColumn
*/
class ResultColumnList extends QueryTreeNodeVector<ResultColumn>
{
/* Is this the ResultColumnList for an index row? */
protected boolean indexRow;
protected long conglomerateId;
int orderBySelect = 0; // the number of result columns pulled up
// from ORDERBY list
/*
* A comment on 'orderBySelect'. When we encounter a SELECT .. ORDER BY
* statement, the columns (or expressions) in the ORDER BY clause may
* or may not have been explicitly mentioned in the SELECT column list.
* If the columns were NOT explicitly mentioned in the SELECT column
* list, then the parsing of the ORDER BY clause implicitly generates
* them into the result column list, because we'll need to have those
* columns present at execution time in order to sort by them. Those
* generated columns are added to the *end* of the ResultColumnList, and
* we keep track of the *number* of those columns in 'orderBySelect',
* so we can tell whether we are looking at a generated column by seeing
* whether its position in the ResultColumnList is in the last
* 'orderBySelect' number of columns. If the SELECT .. ORDER BY
* statement uses the "*" token to select all the columns from a table,
* then during ORDER BY parsing we redundantly generate the columns
* mentioned in the ORDER BY clause into the ResultColumnlist, but then
* later in getOrderByColumnToBind we determine that these are
* duplicates and we take them back out again.
*/
/*
** Is this ResultColumnList for a FromBaseTable for an index
** that is to be updated?
*/
protected boolean forUpdate;
// Is a count mismatch allowed - see set/get methods for details.
private boolean countMismatchAllowed;
// Number of RCs in this RCL at "init" time, before additional
// ones were added internally.
private int initialListSize = 0;
ResultColumnList(ContextManager cm) {
super(ResultColumn.class, cm);
}
/**
* Add a ResultColumn (at this point, ResultColumn or
* AllResultColumn) to the list
*
* @param resultColumn The ResultColumn to add to the list
*/
void addResultColumn(ResultColumn resultColumn)
{
/* Vectors are 0-based, ResultColumns are 1-based */
resultColumn.setVirtualColumnId(size() + 1);
addElement(resultColumn);
}
/**
* Append a given ResultColumnList to this one, resetting the virtual
* column ids in the appended portion.
*
* @param resultColumns The ResultColumnList to be appended
* @param destructiveCopy Whether or not this is a descructive copy
* from resultColumns
*/
void appendResultColumns(ResultColumnList resultColumns,
boolean destructiveCopy)
{
int oldSize = size();
int newID = oldSize + 1;
/*
** Set the virtual column ids in the list being appended.
** Vectors are zero-based, and virtual column ids are one-based,
** so the new virtual column ids start at the original size
** of this list, plus one.
*/
for (ResultColumn rc : resultColumns)
{
/* ResultColumns are 1-based */
rc.setVirtualColumnId(newID);
newID++;
}
if (destructiveCopy)
{
destructiveAppend(resultColumns);
}
else
{
nondestructiveAppend(resultColumns);
}
}
/**
* Get a ResultColumn from a column position (1-based) in the list
*
* @param position The ResultColumn to get from the list (1-based)
*
* @return the column at that position.
*/
ResultColumn getResultColumn(int position)
{
/*
** First see if it falls in position x. If not,
** search the whole shebang
*/
if (position <= size())
{
// this wraps the cast needed,
// and the 0-based nature of the Vectors.
ResultColumn rc = elementAt(position-1);
if (rc.getColumnPosition() == position)
{
return rc;
}
}
/*
** Check each column
*/
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn rc = elementAt(index);
if (rc.getColumnPosition() == position)
{
return rc;
}
}
return null;
}
/**
* Take a column position and a ResultSetNode and find the ResultColumn
* in this RCL whose source result set is the same as the received
* RSN and whose column position is the same as the received column
* position.
*
* @param colNum The column position (w.r.t rsn) for which we're searching
* @param rsn The result set node for which we're searching.
* @return The ResultColumn in this RCL whose source is column colNum
* in result set rsn. That ResultColumn's position w.r.t to this RCL
* is also returned via the whichRC parameter. If no match is found,
* return null and leave whichRC untouched.
*/
ResultColumn getResultColumn(int colNum, ResultSetNode rsn,
int [] whichRC) throws StandardException
{
if (colNum == -1)
return null;
int [] crColNum = new int[] { -1 };
for (int index = size() - 1; index >= 0; index--)
{
ResultColumn rc = elementAt(index);
if (!(rc.getExpression() instanceof ColumnReference))
{
// If the rc's expression isn't a column reference then
// it can't be pointing to rsn, so just skip it.
continue;
}
ColumnReference colRef = (ColumnReference)rc.getExpression();
if ((rsn == colRef.getSourceResultSet(crColNum)) &&
(crColNum[0] == colNum))
{
// Found a match.
whichRC[0] = index+1;
return rc;
}
}
return null;
}
/**
* Get a ResultColumn from a column position (1-based) in the list,
* null if out of range (for order by).
*
* @param position The ResultColumn to get from the list (1-based)
*
* @return the column at that position, null if out of range
*/
ResultColumn getOrderByColumn(int position)
{
// this wraps the cast needed, and the 0-based nature of the Vectors.
if (position == 0)
return null;
return getResultColumn(position);
}
/**
* Get a ResultColumn that matches the specified columnName and
* mark the ResultColumn as being referenced.
*
* @param columnName The ResultColumn to get from the list
*
* @return the column that matches that name.
*/
ResultColumn getResultColumn(String columnName)
{
return getResultColumn( columnName, true );
}
/**
* Get a ResultColumn that matches the specified columnName. If requested
* to, mark the column as referenced.
*
* @param columnName The ResultColumn to get from the list
* @param markIfReferenced True if we should mark this column as referenced.
*
* @return the column that matches that name.
*/
ResultColumn getResultColumn(String columnName, boolean markIfReferenced )
{
for (ResultColumn resultColumn : this)
{
if (columnName.equals( resultColumn.getName()) )
{
/* Mark ResultColumn as referenced and return it */
if ( markIfReferenced ) { resultColumn.setReferenced(); }
return resultColumn;
}
}
return null;
}
/**
* Return a result column, if any found, which contains in its
* expression/&#123;VCN,CR&#125; chain a result column with the given
* columnNumber from a FromTable with the given tableNumber.
* <p/>
* Used by the optimizer preprocess phase when it is flattening queries,
* which has to use the pair &#123;table number, column number&#125; to
* uniquely distinguish the column desired in situations where the same
* table may appear multiple times in the queries with separate correlation
* names, and/or column names from different tables may be the same (hence
* looking up by column name will not always work), cf DERBY-4679.
* <p/>
* {@code columnName} is used to assert that we find the right column.
* If we found a match on (tn, cn) but columnName is wrong, return null.
* Once we trust table numbers and column numbers to always be correct,
* cf. DERBY-4695, we could remove this parameter.
*
* @param tableNumber the table number to look for
* @param columnNumber the column number to look for
* @param columnName name of the desired column
*/
public ResultColumn getResultColumn(int tableNumber,
int columnNumber,
String columnName)
{
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn resultColumn = elementAt(index);
ResultColumn rc = resultColumn;
while (rc != null) {
ValueNode exp = rc.getExpression();
if (exp instanceof VirtualColumnNode) {
VirtualColumnNode vcn = (VirtualColumnNode)exp;
ResultSetNode rsn = vcn.getSourceResultSet();
if (rsn instanceof FromTable) {
FromTable ft = (FromTable)rsn;
if (ft.getTableNumber() == tableNumber) {
// We have the right table, now try to match the
// column number. Looking at a join, for a base
// table participant, we will find the correct
// column position in the
// JOIN's ColumnDescriptor. Normally, we could just
// call rc.getColumnPosition, but this doesn't work
// if we have a join with a subquery participant
// (it would give us the virtualColumnId one level
// too high up, since the column descriptor is null
// in that case inside a JOIN's RC.
//
// If FromTable is a FromSubquery we need to look
// at the JOIN RC's source column to match the
// table column number. However, at that level, the
// table number would be that of the underlying
// SELECT (for example), rather than the
// FromSubquery's, so we need to match the table
// number one level above, cf the test cases in
// JoinTest#testDerby_4679 which have subqueries.
ColumnDescriptor cd = rc.getTableColumnDescriptor();
if (SanityManager.DEBUG) {
SanityManager.ASSERT(
cd != null || ft instanceof FromSubquery);
}
if ( (cd != null && cd.getPosition() ==
columnNumber) ||
(vcn.getSourceColumn().getColumnPosition() ==
columnNumber) ) {
// Found matching (t,c) within this top
// resultColumn. Now do sanity check that column
// name is correct. Remove when DERBY-4695 is
// fixed.
if (columnName.equals(
vcn.getSourceColumn().getName())) {
resultColumn.setReferenced();
return resultColumn;
} else {
if (SanityManager.DEBUG) {
SanityManager.ASSERT(
false,
"wrong (tn,cn) for column " +
columnName +
" found: this pair points to " +
vcn.getSourceColumn().getName());
}
// Fall back on column name based lookup,
// cf. DERBY-4679. See ColumnReference#
// remapColumnReferencesToExpressions
return null;
}
} else {
rc = vcn.getSourceColumn();
}
} else {
rc = vcn.getSourceColumn();
}
} else {
rc = null;
}
} else if (exp instanceof ColumnReference) {
ColumnReference cr = (ColumnReference)exp;
if (cr.getTableNumber() == tableNumber &&
cr.getColumnNumber() == columnNumber) {
// Found matching (t,c) within this top resultColumn
resultColumn.setReferenced();
return resultColumn;
} else {
rc = null;
}
} else {
if (SanityManager.DEBUG) {
SanityManager.ASSERT(
exp instanceof BaseColumnNode,
"expected BaseColumnNode, found: " +
exp.getClass());
}
rc = null;
}
}
}
return null;
}
/**
* Get a ResultColumn that matches the specified columnName and
* mark the ResultColumn as being referenced.
*
* @param columnsTableName Qualifying name for the column
* @param columnName The ResultColumn to get from the list
*
* @return the column that matches that name.
*/
ResultColumn getResultColumn(String columnsTableName, String columnName)
{
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn resultColumn = elementAt(index);
/* If the column's table name is non-null, then we have found a match
* only if the RC's table name is non-null and the same as the
* the CR's table name.
*/
if (columnsTableName != null)
{
if (resultColumn.getTableName() == null)
{
continue;
}
if (! columnsTableName.equals(resultColumn.getTableName()))
{
continue;
}
}
if (columnName.equals( resultColumn.getName()) )
{
/* Mark ResultColumn as referenced and return it */
resultColumn.setReferenced();
return resultColumn;
}
}
return null;
}
/**
* Get a ResultColumn that matches the specified columnName and
* mark the ResultColumn as being referenced.
* NOTE - this flavor enforces no ambiguity (at most 1 match)
* Only FromSubquery needs to call this flavor since
* it can have ambiguous references in its own list.
*
* @param cr The ColumnReference to resolve
* @param exposedTableName Exposed table name for FromTable
* @param considerGeneratedColumns Also consider columns that are generated.
* One example of this is group by where columns are added to the select list
* if they are referenced in the group by but are not present in the select
* list.
* @return the column that matches that name.
*
* @exception StandardException Thrown on error
*/
ResultColumn getAtMostOneResultColumn(
ColumnReference cr,
String exposedTableName,
boolean considerGeneratedColumns)
throws StandardException
{
ResultColumn retRC = null;
String columnName = cr.getColumnName();
for (ResultColumn resultColumn : this)
{
if (columnName.equals( resultColumn.getName()))
{
if (resultColumn.isGenerated() && !considerGeneratedColumns) {
continue;
}
/* We should get at most 1 match */
if (retRC != null)
{
throw StandardException.newException(SQLState.LANG_AMBIGUOUS_COLUMN_NAME_IN_TABLE,
columnName, exposedTableName);
}
/* Mark ResultColumn as referenced and return it */
resultColumn.setReferenced();
retRC = resultColumn;
}
}
return retRC;
}
/**
* Return true if some columns in this list are updatable.
*
* @return true if any column in list is updatable, else false
*/
boolean columnsAreUpdatable()
{
for (ResultColumn rc : this)
{
if (rc.isUpdatable())
return true;
}
return false;
}
/**
* For order by column bind, get a ResultColumn that matches the specified
* columnName.
*
* This method is called during bind processing, in the special
* "bind the order by" call that is made by CursorNode.bindStatement().
* The OrderByList has a special set of bind processing routines
* that analyzes the columns in the ORDER BY list and verifies that
* each column is one of:
* - a direct reference to a column explicitly mentioned in
* the SELECT list
* - a direct reference to a column implicitly mentioned as "SELECT *"
* - a direct reference to a column "pulled up" into the result
* column list
* - or a valid and fully-bound expression ("c+2", "YEAR(hire_date)", etc.)
*
* At this point in the processing, it is possible that we'll find
* the column present in the RCL twice: once because it was pulled
* up during statement compilation, and once because it was added
* when "SELECT *" was expanded into the table's actual column list.
* If we find such a duplicated column, we can, and do, remove the
* pulled-up copy of the column and point the OrderByColumn
* to the actual ResultColumn from the *-expansion.
*
* Note that the association of the OrderByColumn with the
* corresponding ResultColumn in the RCL occurs in
* OrderByColumn.resolveAddedColumn.
*
* @param columnName The ResultColumn to get from the list
* @param tableName The table name on the OrderByColumn, if any
* @param tableNumber The tableNumber corresponding to the FromTable with the
* exposed name of tableName, if tableName != null.
* @param obc The OrderByColumn we're binding.
*
* @return the column that matches that name.
* @exception StandardException thrown on ambiguity
*/
ResultColumn getOrderByColumnToBind(
String columnName,
TableName tableName,
int tableNumber,
OrderByColumn obc)
throws StandardException
{
int size = size();
ResultColumn retVal = null, resultColumn;
for (int index = 0; index < size; index++)
{
resultColumn = elementAt(index);
/* The order by column is qualified, then it is okay to consider
* this RC if:
* o The RC is qualified and the qualifiers on the order by column
* and the RC are equal().
* o The RC is not qualified, but its expression is a ColumnReference
* from the same table (as determined by the tableNumbers).
*/
boolean columnNameMatches;
if (tableName != null)
{
ValueNode rcExpr = resultColumn.getExpression();
if (! (rcExpr instanceof ColumnReference))
continue;
ColumnReference cr = (ColumnReference) rcExpr;
if( (! tableName.equals( cr.getQualifiedTableName())) && tableNumber != cr.getTableNumber())
{
continue;
}
columnNameMatches =
columnName.equals( resultColumn.getSourceColumnName() );
}
else
{
columnNameMatches =
resultColumn.columnNameMatches(columnName);
}
/* We finally got past the qualifiers, now see if the column
* names are equal. If they are, then we appear to have found
* our order by column. If we find our order by column multiple
* times, make sure that they are truly duplicates, otherwise
* we have an ambiguous situation. For example, the query
* SELECT b+c AS a, d+e AS a FROM t ORDER BY a
* is ambiguous because we don't know which "a" is meant. But
* SELECT t.a, t.* FROM t ORDER BY a
* is not ambiguous, even though column "a" is selected twice.
* If we find our ORDER BY column at the end of the
* SELECT column list, in the last 'orderBySelect' number
* of columns, then this column was not explicitly mentioned
* by the user in their SELECT column list, but was implicitly
* added by the parsing of the ORDER BY clause, and it
* should be removed from the ResultColumnList and returned
* to the caller.
*/
if (columnNameMatches)
{
if (retVal == null)
{
retVal = resultColumn;
}
else if (! retVal.isEquivalent(resultColumn))
{
throw StandardException.newException(SQLState.LANG_DUPLICATE_COLUMN_FOR_ORDER_BY, columnName);
}
else if (index >= size - orderBySelect)
{// remove the column due to pullup of orderby item
removeElement(resultColumn);
decOrderBySelect();
obc.clearAddedColumnOffset();
collapseVirtualColumnIdGap(
resultColumn.getColumnPosition());
break;
}
}
}
return retVal;
}
/**
* Adjust virtualColumnId values due to result column removal
*
* This method is called when a duplicate column has been detected and
* removed from the list. We iterate through each of the other columns
* in the list and notify them of the column removal so they can adjust
* their virtual column id if necessary.
*
* @param gap id of the column which was just removed.
*/
private void collapseVirtualColumnIdGap(int gap)
{
for (ResultColumn rc : this) {
rc.collapseVirtualColumnIdGap(gap);
}
}
/**
* For order by, get a ResultColumn that matches the specified
* columnName.
*
* This method is called during pull-up processing, at the very
* start of bind processing, as part of
* OrderByList.pullUpOrderByColumns. Its job is to figure out
* whether the provided column (from the ORDER BY list) already
* exists in the ResultColumnList or not. If the column does
* not exist in the RCL, we return NULL, which signifies that
* a new ResultColumn should be generated and added ("pulled up")
* to the RCL by our caller.
*
* Note that at this point in the processing, we should never
* find this column present in the RCL multiple times; if the
* column is already present in the RCL, then we don't need to,
* and won't, pull a new ResultColumn up into the RCL.
*
* If the caller specified "SELECT *", then the RCL at this
* point contains a special AllResultColumn object. This object
* will later be expanded and replaced by the actual set of
* columns in the table, but at this point we don't know what
* those columns are, so we may pull up an OrderByColumn
* which duplicates a column in the *-expansion; such
* duplicates will be removed at the end of bind processing
* by OrderByList.bindOrderByColumns.
*
* @param columnName The ResultColumn to get from the list
* @param tableName The table name on the OrderByColumn, if any
*
* @return the column that matches that name, or NULL if pull-up needed
* @exception StandardException thrown on ambiguity
*/
ResultColumn findResultColumnForOrderBy(
String columnName, TableName tableName)
throws StandardException
{
int size = size();
ResultColumn retVal = null, resultColumn;
for (int index = 0; index < size; index++)
{
resultColumn = elementAt(index);
// We may be checking on "ORDER BY T.A" against "SELECT *".
// exposedName will not be null and "*" will not have an expression
// or tablename.
// We may be checking on "ORDER BY T.A" against "SELECT T.B, T.A".
boolean columnNameMatches;
if (tableName != null)
{
ValueNode rcExpr = resultColumn.getExpression();
if (rcExpr == null || ! (rcExpr instanceof ColumnReference))
{
continue;
}
ColumnReference cr = (ColumnReference) rcExpr;
if( ! tableName.equals( cr.getQualifiedTableName()))
{
continue;
}
columnNameMatches =
columnName.equals( resultColumn.getSourceColumnName() );
}
else
columnNameMatches =
resultColumn.columnNameMatches(columnName);
/* We finally got past the qualifiers, now see if the column
* names are equal.
*/
if (columnNameMatches)
{
if (retVal == null)
{
retVal = resultColumn;
}
else if (! retVal.isEquivalent(resultColumn))
{
throw StandardException.newException(SQLState.LANG_DUPLICATE_COLUMN_FOR_ORDER_BY, columnName);
}
else if (index >= size - orderBySelect)
{
if (SanityManager.DEBUG)
SanityManager.THROWASSERT(
"Unexpectedly found ORDER BY column '" +
columnName + "' pulled up at position " +index);
}
}
}
return retVal;
}
/**
* Copy the result column names from the given ResultColumnList
* to this ResultColumnList. This is useful for insert-select,
* where the columns being inserted into may be different from
* the columns being selected from. The result column list for
* an insert is supposed to have the column names being inserted
* into.
*
* @param nameList The ResultColumnList from which to copy
* the column names
*/
void copyResultColumnNames(ResultColumnList nameList)
{
/* List checking is done during bind(). Lists should be the
* same size when we are called.
*/
if (SanityManager.DEBUG)
{
if ( (! countMismatchAllowed) &&
visibleSize() != nameList.visibleSize() )
{
SanityManager.THROWASSERT(
"The size of the 2 lists is expected to be the same. " +
"visibleSize() = " + visibleSize() +
", nameList.visibleSize() = " + nameList.visibleSize());
}
}
int size =
(countMismatchAllowed) ? nameList.visibleSize() : visibleSize();
for (int index = 0; index < size; index++)
{
ResultColumn thisResultColumn = elementAt(index);
ResultColumn nameListResultColumn = nameList.elementAt(index);
thisResultColumn.setName(nameListResultColumn.getName());
thisResultColumn.setNameGenerated(nameListResultColumn.isNameGenerated());
}
}
/**
* Bind the expressions in this ResultColumnList. This means binding
* the expression under each ResultColumn node.
*
* @param fromList The FROM list for the query this
* expression is in, for binding columns.
* @param subqueryList The subquery list being built as we find SubqueryNodes
* @param aggregates The aggregate list being built as we find AggregateNodes
*
* @exception StandardException Thrown on error
*/
void bindExpressions(
FromList fromList, SubqueryList subqueryList, List<AggregateNode> aggregates)
throws StandardException
{
/* First we expand the *'s in the result column list */
expandAllsAndNameColumns(fromList);
/* Now we bind each result column */
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn vn = elementAt(index);
vn = vn.bindExpression(fromList, subqueryList, aggregates);
setElementAt(vn, index);
}
}
/**
* Bind the result columns to the expressions that live under them.
* All this does is copy the datatype information to from each expression
* to each result column. This is useful for SELECT statements, where
* the result type of each column is the type of the column's expression.
*
* @exception StandardException Thrown on error
*/
void bindResultColumnsToExpressions()
throws StandardException
{
for (ResultColumn rc : this)
{
rc.bindResultColumnToExpression();
}
}
/**
* Bind the result columns by their names. This is useful for GRANT and REVOKE statements
* like "GRANT SELECT ON t(c1,c1,c3) TO george", where the user specified a column list.
* This method does not check for duplicate column names.
*
* @param targetTableDescriptor The descriptor for the table
*
* @exception StandardException Thrown on error
*/
void bindResultColumnsByName(TableDescriptor targetTableDescriptor)
throws StandardException
{
int size = size();
for (int index = 0; index < size; index++)
{
elementAt(index).bindResultColumnByName(
targetTableDescriptor,
index + 1
);
}
} // end of bindResultColumnsByName( TableDescriptor)
/**
* Bind the result columns by their names. This is useful for update, grant, and revoke
* statements, and for INSERT statements like "insert into t (a, b, c)
* values (1, 2, 3)" where the user specified a column list.
* If the statment is an insert or update verify that the result column list does not contain any duplicates.
* NOTE: We pass the ResultColumns position in the ResultColumnList so
* that the VirtualColumnId gets set.
*
* @param targetTableDescriptor The descriptor for the table being
* updated or inserted into
* @param statement DMLStatementNode containing this list, null if no duplicate checking is to be done
*
* @return A FormatableBitSet representing the set of columns with respect to the table
*
* @exception StandardException Thrown on error
*/
FormatableBitSet bindResultColumnsByName(TableDescriptor targetTableDescriptor,
DMLStatementNode statement)
throws StandardException
{
int size = size();
FormatableBitSet columnBitSet = new FormatableBitSet( targetTableDescriptor.getNumberOfColumns());
for (int index = 0; index < size; index++)
{
ResultColumn rc = elementAt(index);
rc.bindResultColumnByName(
targetTableDescriptor,
index + 1
);
int colIdx = rc.getColumnPosition() - 1;
if( SanityManager.DEBUG)
SanityManager.ASSERT( colIdx >= 0 && colIdx < targetTableDescriptor.getNumberOfColumns(),
"Invalid column position found for " + rc.getName());
/* Verify that this column's name is unique within the list if requested */
if( statement != null && columnBitSet.isSet( colIdx))
{
String colName = rc.getName();
if (statement instanceof UpdateNode)
{
throw StandardException.newException(SQLState.LANG_DUPLICATE_COLUMN_NAME_UPDATE, colName);
}
else
{
throw StandardException.newException(SQLState.LANG_DUPLICATE_COLUMN_NAME_INSERT, colName);
}
}
columnBitSet.set( colIdx);
}
return columnBitSet;
}
/**
* Bind the result columns by their names. This is useful for update
* VTI statements, and for INSERT statements like "insert into new t() (a, b, c)
* values (1, 2, 3)" where the user specified a column list.
* Also, verify that the result column list does not contain any duplicates.
* NOTE: We pass the ResultColumns position in the ResultColumnList so
* that the VirtualColumnId gets set.
*
* @param fullRCL The full RCL for the target table
* @param statement DMLStatementNode containing this list
*
* @exception StandardException Thrown on error
*/
void bindResultColumnsByName(ResultColumnList fullRCL,
FromVTI targetVTI,
DMLStatementNode statement)
throws StandardException
{
int size = size();
HashSet<String> seenNames = new HashSet<String>(size + 2, 0.999f);
for (int index = 0; index < size; index++)
{
ResultColumn matchRC;
ResultColumn rc = elementAt(index);
/* Verify that this column's name is unique within the list */
String colName = rc.getName();
boolean alreadySeen = !seenNames.add(colName);
if (alreadySeen)
{
if (SanityManager.DEBUG)
{
SanityManager.ASSERT((statement instanceof UpdateNode) ||
(statement instanceof InsertNode),
"statement is expected to be instanceof UpdateNode or InsertNode");
}
if (statement instanceof UpdateNode)
{
throw StandardException.newException(SQLState.LANG_DUPLICATE_COLUMN_NAME_UPDATE, colName);
}
else
{
throw StandardException.newException(SQLState.LANG_DUPLICATE_COLUMN_NAME_INSERT, colName);
}
}
matchRC = fullRCL.getResultColumn(null, rc.getName());
if (matchRC == null)
{
throw StandardException.newException(SQLState.LANG_COLUMN_NOT_FOUND_IN_TABLE,
rc.getName(),
targetVTI.getMethodCall().getJavaClassName());
}
/* We have a match. We need to create a dummy ColumnDescriptor
* since calling code expects one to get column info.
*/
ColumnDescriptor cd = new ColumnDescriptor(
rc.getName(),
matchRC.getVirtualColumnId(),
matchRC.getType(),
null,
null,
(TableDescriptor) null,
null,
0, 0);
rc.setColumnDescriptor(null, cd);
rc.setVirtualColumnId(index + 1);
}
}
/**
* Bind the result columns by ordinal position. This is useful for
* INSERT statements like "insert into t values (1, 2, 3)", where the
* user did not specify a column list.
*
* @param targetTableDescriptor The descriptor for the table being
* inserted into
*
* @exception StandardException Thrown on error
*/
void bindResultColumnsByPosition(TableDescriptor targetTableDescriptor)
throws StandardException
{
int size = size();
for (int index = 0; index < size; index++)
{
/*
** Add one to the iterator index, because iterator indexes start at zero,
** and column numbers start at one.
*/
elementAt(index).bindResultColumnByPosition(
targetTableDescriptor,
index + 1);
}
}
/**
* Preprocess the expression trees under the RCL.
* We do a number of transformations
* here (including subqueries, IN lists, LIKE and BETWEEN) plus
* subquery flattening.
* NOTE: This is done before the outer ResultSetNode is preprocessed.
*
* @param numTables Number of tables in the DML Statement
* @param outerFromList FromList from outer query block
* @param outerSubqueryList SubqueryList from outer query block
* @param outerPredicateList PredicateList from outer query block
*
* @exception StandardException Thrown on error
*/
void preprocess(int numTables,
FromList outerFromList,
SubqueryList outerSubqueryList,
PredicateList outerPredicateList)
throws StandardException
{
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn resultColumn = elementAt(index);
setElementAt(resultColumn.preprocess(numTables,
outerFromList,
outerSubqueryList,
outerPredicateList),
index);
}
}
/**
Verify that all the result columns have expressions that
are storable for them. Check versus the given ResultColumnList.
@exception StandardException Thrown on error
*/
void checkStorableExpressions(ResultColumnList toStore)
throws StandardException
{
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn otherRC = toStore.elementAt(index);
elementAt(index).checkStorableExpression(otherRC);
}
}
/**
Return an array holding the 0 based heap offsets of
the StreamStorable columns in this ResultColumnList.
This returns null if this list does not contain any
StreamStorableColumns. The list this returns does not
contain duplicates. This should only be used for
a resultColumnList the refers to a single heap
such as the target for an Insert, Update or Delete.
@param heapColCount the number of heap columns
@exception StandardException Thrown on error
*/
int[] getStreamStorableColIds(int heapColCount) throws StandardException
{
//@#$
//System.out.println("getStreamStorableColids");
int ssCount = 0;
boolean[] isSS = new boolean[heapColCount];//Should be table length.
for (ResultColumn rc : this)
{
if (rc.getTypeId().streamStorable())
{
//System.out.println(" streamStorable=true");
ColumnDescriptor cd = rc.getTableColumnDescriptor();
isSS[cd.getPosition()-1] = true;
}
}
for (int ix=0;ix<isSS.length;ix++) if (isSS[ix]) ssCount++;
if (ssCount==0)return null;
int[] result = new int[ssCount];
int resultOffset=0;
for (int heapOffset=0;heapOffset<isSS.length;heapOffset++)
{
if (isSS[heapOffset])
result[resultOffset++]=heapOffset;
}
return result;
}
/**
Verify that all the result columns have expressions that
are storable for them. Check versus the expressions under the
ResultColumns.
@exception StandardException Thrown on error
*/
void checkStorableExpressions()
throws StandardException
{
for (ResultColumn rc : this)
{
rc.checkStorableExpression();
}
}
/**
* Generate the code to place the columns' values into
* a row variable named "r". This wrapper is here
* rather than in ResultColumn, because that class does
* not know about the position of the columns in the list.
*
* @exception StandardException Thrown on error
*/
@Override
void generate(ActivationClassBuilder acb, MethodBuilder mb)
throws StandardException
{
generateCore(acb, mb, false);
}
/**
* Generate the code to place the columns' values into
* a row variable named "r". This wrapper is here
* rather than in ResultColumn, because that class does
* not know about the position of the columns in the list.
*
* @exception StandardException Thrown on error
*/
void generateNulls(ActivationClassBuilder acb,
MethodBuilder mb)
throws StandardException
{
generateCore(acb, mb, true);
}
/**
* Generate the code to place the columns' values into
* a row variable named "r". This wrapper is here
* rather than in ResultColumn, because that class does
* not know about the position of the columns in the list.
*
* This is the method that does the work.
*/
void generateCore(ExpressionClassBuilder acb,
MethodBuilder mb,
boolean genNulls)
throws StandardException
{
// generate the function and initializer:
// private ExecRow fieldX;
// In the constructor:
// fieldX = getExecutionFactory().getValueRow(# cols);
// private ExecRow exprN()
// {
// fieldX.setColumn(1, col(1).generateColumn(ps)));
// ... and so on for each column ...
// return fieldX;
// }
// static Method exprN = method pointer to exprN;
// this sets up the method and the static field.
MethodBuilder userExprFun = acb.newUserExprFun();
generateEvaluatedRow( acb, userExprFun, genNulls, false );
// what we return is the access of the field, i.e. the pointer to the method.
acb.pushMethodReference(mb, userExprFun);
}
/**
* <p>
* Generate the code for a method (userExprFun) which creates a row
* and, column by column, stuffs it with the evaluated
* expressions of our ResultColumns. The method returns the
* stuffed row.
* </p>
*
* This is the method that does the work.
*/
void generateEvaluatedRow
(
ExpressionClassBuilder acb,
MethodBuilder userExprFun,
boolean genNulls,
boolean forMatchingClause
)
throws StandardException
{
// generate the function and initializer:
// private ExecRow fieldX;
// In the constructor:
// fieldX = getExecutionFactory().getValueRow(# cols);
// private ExecRow exprN()
// {
// fieldX.setColumn(1, col(1).generateColumn(ps)));
// ... and so on for each column ...
// return fieldX;
// }
// static Method exprN = method pointer to exprN;
/* Declare the field */
LocalField field = acb.newFieldDeclaration(Modifier.PRIVATE, ClassName.ExecRow);
// Generate the code to create the row in the constructor
genCreateRow(acb, field, "getValueRow", ClassName.ExecRow, size());
ResultColumn rc;
int size = size();
MethodBuilder cb = acb.getConstructor();
for (int index = 0; index < size; index++)
{
// generate statements of the form
// fieldX.setColumn(columnNumber, (DataValueDescriptor) columnExpr);
// and add them to exprFun.
rc = elementAt(index);
/* If we are not generating nulls, then we can skip this RC if
* it is simply propagating a column from the source result set.
*/
if (!genNulls)
{
ValueNode sourceExpr = rc.getExpression();
if ( sourceExpr instanceof VirtualColumnNode && ! ( ((VirtualColumnNode) sourceExpr).getCorrelated()) )
{
continue;
}
//DERBY-4631 - For INNER JOINs and LEFT OUTER
// JOINs, Derby retrieves the join column values
// from the left table's join column. But for
// RIGHT OUTER JOINs, the join column's value
// will be picked up based on following logic.
//1)if the left table's column value is null
// then pick up the right table's column's value.
//2)If the left table's column value is non-null,
// then pick up that value
if (rc.getJoinResultSet() != null) {
//We are dealing with a join column for
// RIGHT OUTER JOIN with USING/NATURAL eg
// select c from t1 right join t2 using (c)
//We are talking about column c as in "select c"
ResultColumnList jnRCL =
rc.getJoinResultSet().getResultColumns();
int joinResultSetNumber =
rc.getJoinResultSet().getResultSetNumber();
//We need to know the column positions of left
// table's join column and right table's join
// column to generate the code explained above
int virtualColumnIdRightTable = -1;
int virtualColumnIdLeftTable = -1;
for (ResultColumn joinColumn : jnRCL) {
if (joinColumn.getName().equals(rc.getUnderlyingOrAliasName())) {
if (joinColumn.isRightOuterJoinUsingClause())
virtualColumnIdRightTable = joinColumn.getVirtualColumnId();
else
virtualColumnIdLeftTable = joinColumn.getVirtualColumnId();
}
}
userExprFun.getField(field); // instance
userExprFun.push(index + 1); // arg1
String resultTypeName =
getTypeCompiler(
DataTypeDescriptor.getBuiltInDataTypeDescriptor(
Types.BOOLEAN).getTypeId()).interfaceName();
String receiverType = ClassName.DataValueDescriptor;
//Our plan is to generate DERBY-4631
// if(lefTablJoinColumnValue is null)
// then
// use rightTablJoinColumnValue
// else
// use lefTablJoinColumnValue
//Following will generate
// if(lefTablJoinColumnValue is null)
acb.pushColumnReference(userExprFun, joinResultSetNumber,
virtualColumnIdLeftTable);
userExprFun.cast(rc.getTypeCompiler().interfaceName());
userExprFun.cast(receiverType);
userExprFun.callMethod(VMOpcode.INVOKEINTERFACE, (String) null,
"isNullOp",resultTypeName, 0);
//Then call generateExpression on left Table's column
userExprFun.cast(ClassName.BooleanDataValue);
userExprFun.push(true);
userExprFun.callMethod(
VMOpcode.INVOKEINTERFACE, (String) null, "equals", "boolean", 1);
//Following will generate
// then
// use rightTablJoinColumnValue
userExprFun.conditionalIf();
acb.pushColumnReference(userExprFun, joinResultSetNumber,
virtualColumnIdRightTable);
userExprFun.cast(rc.getTypeCompiler().interfaceName());
//Following will generate
// else
// use lefTablJoinColumnValue
userExprFun.startElseCode();
acb.pushColumnReference(userExprFun, joinResultSetNumber,
virtualColumnIdLeftTable);
userExprFun.cast(rc.getTypeCompiler().interfaceName());
userExprFun.completeConditional();
userExprFun.cast(ClassName.DataValueDescriptor);
userExprFun.callMethod(
VMOpcode.INVOKEINTERFACE, ClassName.Row, "setColumn", "void", 2);
continue;
}
if ( !forMatchingClause )
{
if (sourceExpr instanceof ColumnReference && ! ( ((ColumnReference) sourceExpr).getCorrelated()))
{
continue;
}
}
}
// row add is 1-based, and iterator index is 0-based
if (SanityManager.DEBUG)
{
if (index + 1 != rc.getVirtualColumnId())
{
SanityManager.THROWASSERT(
"VirtualColumnId (" +
rc.getVirtualColumnId() +
") does not agree with position within Vector (" +
(index + 1) +
")");
}
}
//
// Generated columns should be populated after the base row because
// the generation clauses may refer to base columns that have to be filled
// in first. Population of generated columns is done in another
// method, which (like CHECK CONSTRAINTS) is explicitly called by
// InsertResultSet and UpdateResultSet.
//
// For LEFT JOINs, we may need to stuff a NULL into the generated column slot,
// just as we do for non-generated columns in a LEFT JOIN. We look at the source
// expression for the ResultColumn to determine whether this ResultColumnList
// represents an INSERT/UPDATE vs. a SELECT. If this ResultColumnList represents a
// LEFT JOIN, then the source expression will be a VirtualColumnNode.
// See DERBY-6346.
//
if ( rc.hasGenerationClause() )
{
ValueNode expr = rc.getExpression();
if ( (expr != null) && !(expr instanceof VirtualColumnNode) )
{
continue;
}
}
// we need the expressions to be Columns exactly.
/* SPECIAL CASE: Expression is a non-null constant.
* Generate the setColumn() call in the constructor
* so that it will only be executed once per instantiation.
*
* Increase the statement counter in constructor. Code size in
* constructor can become too big (more than 64K) for Java compiler
* to handle (beetle 4293). We set constant columns in other
* methods if constructor has too many statements already.
*/
if ( (! genNulls) &&
(rc.getExpression() instanceof ConstantNode) &&
! ((ConstantNode) rc.getExpression()).isNull() &&
! cb.statementNumHitLimit(1))
{
cb.getField(field); // instance
cb.push(index + 1); // first arg;
rc.generateExpression(acb, cb);
cb.cast(ClassName.DataValueDescriptor); // second arg
cb.callMethod(VMOpcode.INVOKEINTERFACE, ClassName.Row, "setColumn", "void", 2);
continue;
}
userExprFun.getField(field); // instance
userExprFun.push(index + 1); // arg1
/* We want to reuse the null values instead of doing a new each time
* if the caller said to generate nulls or the underlying expression
* is a typed null value.
*/
boolean needDVDCast = true;
if (rc.isAutoincrementGenerated())
{
// (com.ibm.db2j.impl... DataValueDescriptor)
// this.getSetAutoincValue(column_number)
userExprFun.pushThis();
userExprFun.push(rc.getColumnPosition());
userExprFun.push(rc.getTableColumnDescriptor().getAutoincInc());
userExprFun.callMethod(VMOpcode.INVOKEVIRTUAL, ClassName.BaseActivation,
"getSetAutoincrementValue", ClassName.DataValueDescriptor, 2);
needDVDCast = false;
}
else if (genNulls ||
((rc.getExpression() instanceof ConstantNode) &&
((ConstantNode) rc.getExpression()).isNull()))
{
userExprFun.getField(field);
userExprFun.push(index + 1);
userExprFun.callMethod(VMOpcode.INVOKEINTERFACE, ClassName.Row, "getColumn",
ClassName.DataValueDescriptor, 1); // the express
acb.generateNullWithExpress(userExprFun, rc.getTypeCompiler(),
rc.getTypeServices().getCollationType());
}
else
{
rc.generateExpression(acb, userExprFun);
}
if (needDVDCast)
userExprFun.cast(ClassName.DataValueDescriptor);
userExprFun.callMethod(VMOpcode.INVOKEINTERFACE, ClassName.Row, "setColumn", "void", 2);
}
userExprFun.getField(field);
userExprFun.methodReturn();
// we are now done modifying userExprFun
userExprFun.complete();
}
/**
* Build an empty row with the size and shape of the ResultColumnList.
*
* @return an empty row of the correct size and shape.
* @exception StandardException Thrown on error
*/
public ExecRow buildEmptyRow()
throws StandardException
{
int columnCount = size();
ExecRow row = getExecutionFactory().getValueRow( columnCount );
int position = 1;
for (ResultColumn rc : this)
{
DataTypeDescriptor dataType = rc.getTypeServices();
DataValueDescriptor dataValue = dataType.getNull();
row.setColumn( position++, dataValue );
}
return row;
}
/**
* Build an empty index row for the given conglomerate.
*
* @return an empty row of the correct size and shape.
* @exception StandardException Thrown on error
*/
public ExecRow buildEmptyIndexRow(TableDescriptor td,
ConglomerateDescriptor cd,
StoreCostController scc,
DataDictionary dd)
throws StandardException
{
ResultColumn rc;
if (SanityManager.DEBUG)
{
if (! cd.isIndex())
{
SanityManager.THROWASSERT("ConglomerateDescriptor expected to be for index: " + cd);
}
}
int[] baseCols = cd.getIndexDescriptor().baseColumnPositions();
ExecRow row = getExecutionFactory().getValueRow(baseCols.length + 1);
for (int i = 0; i < baseCols.length; i++)
{
ColumnDescriptor coldes = td.getColumnDescriptor(baseCols[i]);
DataTypeDescriptor dataType = coldes.getType();
// rc = getResultColumn(baseCols[i]);
// rc = (ResultColumn) at(baseCols[i] - 1);
// dataType = rc.getTypeServices();
DataValueDescriptor dataValue = dataType.getNull();
row.setColumn(i + 1, dataValue );
}
RowLocation rlTemplate = scc.newRowLocationTemplate();
row.setColumn(baseCols.length + 1, rlTemplate);
return row;
}
/**
* Build an {@code ExecRowBuilder} instance that produces a row of the
* same shape as this result column list.
*
* @param referencedCols a bit map that tells which columns in the
* source result set that are used, or {@code null} if all are used
* @param skipPropagatedCols whether to skip virtual columns whose
* source is the immediate child result set
* @return an instance that produces rows of the same shape as this
* result column list
*/
ExecRowBuilder buildRowTemplate(FormatableBitSet referencedCols,
boolean skipPropagatedCols)
throws StandardException
{
int columns = (referencedCols == null) ?
size() : referencedCols.getNumBitsSet();
ExecRowBuilder builder = new ExecRowBuilder(columns, indexRow);
// Get the index of the first column to set in the row template.
int colNum = (referencedCols == null) ? 0 : referencedCols.anySetBit();
for (ResultColumn rc : this) {
ValueNode sourceExpr = rc.getExpression();
if (sourceExpr instanceof CurrentRowLocationNode) {
builder.setColumn(colNum + 1, newRowLocationTemplate());
} else if (skipPropagatedCols &&
sourceExpr instanceof VirtualColumnNode) {
// Skip over those columns whose source is the immediate
// child result set. (No need to generate a wrapper
// for a SQL NULL when we are smart enough not to pass
// that wrapper to the store.)
continue;
} else {
builder.setColumn(colNum + 1, rc.getType());
}
// Get the index of the next column to set in the row template.
if (referencedCols == null) {
colNum++;
} else {
colNum = referencedCols.anySetBit(colNum);
}
}
return builder;
}
/**
* Shorthand for {@code buildRowTemplate(null, false)}.
*/
ExecRowBuilder buildRowTemplate() throws StandardException {
return buildRowTemplate(null, false);
}
/**
* Generate the code to create an empty row in the constructor.
*
* @param acb The ACB.
* @param field The field for the new row.
* @param rowAllocatorMethod The method to call.
* @param rowAllocatorType The row type.
* @param numCols The number of columns in the row.
*
* @exception StandardException Thrown on error
*/
private void genCreateRow(ExpressionClassBuilder acb,
LocalField field,
String rowAllocatorMethod,
String rowAllocatorType,
int numCols)
throws StandardException
{
// Create the row in the constructor
// fieldX = getExecutionFactory().getValueRow(# cols);
MethodBuilder cb = acb.getConstructor();
acb.pushGetExecutionFactoryExpression(cb); // instance
cb.push(numCols);
cb.callMethod(VMOpcode.INVOKEINTERFACE, (String) null,
rowAllocatorMethod, rowAllocatorType, 1);
cb.setField(field);
/* Increase the statement counter in constructor. Code size in
* constructor can become too big (more than 64K) for Java compiler
* to handle (beetle 4293). We set constant columns in other
* methods if constructor has too many statements already.
*/
cb.statementNumHitLimit(1); // ignore return value
}
/**
* Create a row location template of the right type for the source
* conglomerate.
*/
private RowLocation newRowLocationTemplate() throws StandardException {
LanguageConnectionContext lcc = getLanguageConnectionContext();
DataDictionary dd = lcc.getDataDictionary();
int isolationLevel = (dd.getCacheMode() == DataDictionary.DDL_MODE) ?
TransactionController.ISOLATION_READ_COMMITTED :
TransactionController.ISOLATION_NOLOCK;
ConglomerateController cc =
lcc.getTransactionCompile().openConglomerate(
conglomerateId,
false,
0,
TransactionController.MODE_RECORD,
isolationLevel);
try {
return cc.newRowLocationTemplate();
} finally {
cc.close();
}
}
/**
* Make a ResultDescription for use in a ResultSet.
* This is useful when generating/executing a NormalizeResultSet, since
* it can appear anywhere in the tree.
*
* @return A ResultDescription for this ResultSetNode.
*/
ResultColumnDescriptor[] makeResultDescriptors()
{
ResultColumnDescriptor colDescs[] = new ResultColumnDescriptor[size()];
int size = size();
for (int index = 0; index < size; index++)
{
// the ResultColumn nodes are descriptors, so take 'em...
colDescs[index] = getExecutionFactory().getResultColumnDescriptor(((ResultColumnDescriptor) elementAt(index)));
}
return colDescs;
}
/**
* Expand any *'s in the ResultColumnList. In addition, we will guarantee that
* each ResultColumn has a name. (All generated names will be unique across the
* entire statement.)
*
*
* @exception StandardException Thrown on error
*/
void expandAllsAndNameColumns(FromList fromList)
throws StandardException
{
boolean expanded = false;
ResultColumnList allExpansion;
TableName fullTableName;
/* First walk result column list looking for *'s to expand */
for (int index = 0; index < size(); index++)
{
ResultColumn rc = elementAt(index);
if (rc instanceof AllResultColumn)
{
expanded = true;
//fullTableName = ((AllResultColumn) rc).getFullTableName();
TableName temp = rc.getTableNameObject();
if(temp != null) {
String sName = temp.getSchemaName();
String tName = temp.getTableName();
fullTableName = makeTableName(sName,tName);
}
else
fullTableName = null;
allExpansion = fromList.expandAll(fullTableName);
/* Make sure that every column has a name */
allExpansion.nameAllResultColumns();
/* Replace the AllResultColumn with the expanded list.
* We will update the VirtualColumnIds once below.
*/
removeElementAt(index);
for (int inner = 0; inner < allExpansion.size(); inner++)
{
insertElementAt(allExpansion.elementAt(inner), index + inner);
}
// Move the index position to account for the removals and the
// insertions. Should be positioned on the last column in the
// expansion to prevent double processing of the columns.
// DERBY-4410: If the expansion is empty, this will move the
// position one step back because the * was removed and nothing
// was inserted, so all columns to the right of the current
// position have been moved one position to the left. If we
// don't adjust the position, we end up skipping columns.
index += (allExpansion.size() - 1);
// If the rc was a "*", we need to set the initial list size
// to the number of columns that are actually returned to
// the user.
markInitialSize();
}
else
{
/* Make sure that every column has a name */
rc.guaranteeColumnName();
}
}
/* Go back and update the VirtualColumnIds if we expanded any *'s */
if (expanded)
{
int size = size();
for (int index = 0; index < size; index++)
{
/* Vectors are 0-based, VirtualColumnIds are 1-based. */
elementAt(index).setVirtualColumnId(index + 1);
}
}
}
/**
* Generate (unique across the entire statement) column names for those
* ResultColumns in this list which are not named.
*
* @exception StandardException Thrown on error
*/
void nameAllResultColumns()
throws StandardException
{
for (ResultColumn rc : this)
{
rc.guaranteeColumnName();
}
}
/**
** Check whether the column lengths and types of the result columns
** match the expressions under those columns. This is useful for
** INSERT and UPDATE statements. For SELECT statements this method
** should always return true. There is no need to call this for a
** DELETE statement.
** NOTE: We skip over generated columns since they won't have a
** column descriptor.
**
** @return true means all the columns match their expressions,
** false means at least one column does not match its
** expression
*/
boolean columnTypesAndLengthsMatch()
throws StandardException
{
for (ResultColumn resultColumn : this)
{
/* Skip over generated columns */
if (resultColumn.isGenerated())
{
continue;
}
if (! resultColumn.columnTypeAndLengthMatch())
return false;
}
return true;
}
boolean columnTypesAndLengthsMatch(ResultColumnList otherRCL)
throws StandardException
{
boolean retval = true;
/* We check every RC, even after finding 1 that requires
* normalization, because the conversion of constants to
* the appropriate type occurs under this loop.
*/
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn resultColumn = elementAt(index);
ResultColumn otherResultColumn = otherRCL.elementAt(index);
/* Skip over generated columns */
if (resultColumn.isGenerated() || otherResultColumn.isGenerated())
{
continue;
}
if (! resultColumn.columnTypeAndLengthMatch(otherResultColumn))
{
retval = false;
}
}
return retval;
}
/**
* Determine whether this RCL is a No-Op projection of the given RCL.
* It only makes sense to do this if the given RCL is from the child
* result set of the ProjectRestrict that this RCL is from.
*
* @param childRCL The ResultColumnList of the child result set.
*
* @return true if this RCL is a No-Op projection of the given RCL.
*/
boolean nopProjection(ResultColumnList childRCL)
{
/*
** This RCL is a useless projection if each column in the child
** if the same as the column in this RCL. This is impossible
** if the two RCLs have different numbers of columns.
*/
if (this.size() != childRCL.size())
{
return false;
}
/*
** The two lists have the same numbers of elements. Are the lists
** identical? In other words, is the expression in every ResultColumn
** in the PRN's RCL a ColumnReference that points to the corresponding
** column in the child?
*/
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn thisColumn = elementAt(index);
ResultColumn referencedColumn;
/*
** A No-Op projection can point to a VirtualColumnNode or a
** ColumnReference.
*/
if (thisColumn.getExpression() instanceof VirtualColumnNode)
{
referencedColumn =
((VirtualColumnNode) (thisColumn.getExpression())).
getSourceColumn();
}
else if (thisColumn.getExpression() instanceof ColumnReference)
{
referencedColumn =
((ColumnReference) (thisColumn.getExpression())).
getSource();
}
else
{
return false;
}
ResultColumn childColumn = childRCL.elementAt(index);
if (referencedColumn != childColumn)
{
return false;
}
}
return true;
}
/**
* Create a shallow copy of a ResultColumnList and its ResultColumns.
* (All other pointers are preserved.)
* Useful for building new ResultSetNodes during preprocessing.
*
* @return None.
*
* @exception StandardException Thrown on error
*/
ResultColumnList copyListAndObjects()
throws StandardException
{
ResultColumnList newList = new ResultColumnList(getContextManager());
/* Walk the current list - for each ResultColumn in the list, make a copy
* and add it to the new list.
*/
for (ResultColumn origResultColumn: this)
{
newList.addResultColumn(origResultColumn.cloneMe());
}
newList.copyOrderBySelect(this);
return newList;
}
/**
* Remove any columns that may have been added for an order by clause.
* In a query like:
* <pre>select a from t order by b</pre> b is added to the select list
* However in the final projection, after the sort is complete, b will have
* to be removed.
*
*/
void removeOrderByColumns()
{
int idx = size() - 1;
for (int i = 0; i < orderBySelect; i++, idx--) {
removeElementAt(idx);
}
orderBySelect = 0;
}
/**
* Walk the list and replace ResultColumn.expression with a new
* VirtualColumnNode. This is useful when propagating a ResultColumnList
* up the query tree.
* NOTE: This flavor marks all of the underlying RCs as referenced.
*
* @param sourceResultSet ResultSetNode that is source of value
*
* @exception StandardException Thrown on error
*/
void genVirtualColumnNodes(ResultSetNode sourceResultSet,
ResultColumnList sourceResultColumnList)
throws StandardException
{
genVirtualColumnNodes(sourceResultSet, sourceResultColumnList, true);
}
/**
* Walk the list and replace ResultColumn.expression with a new
* VirtualColumnNode. This is useful when propagating a ResultColumnList
* up the query tree.
*
* @param sourceResultSet ResultSetNode that is source of value
* @param markReferenced Whether or not to mark the underlying RCs
* as referenced
*
* @exception StandardException Thrown on error
*/
void genVirtualColumnNodes(ResultSetNode sourceResultSet,
ResultColumnList sourceResultColumnList,
boolean markReferenced)
throws StandardException
{
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn resultColumn = elementAt(index);
/* Vectors are 0-based, VirtualColumnIds are 1-based */
resultColumn.setExpression
(
new VirtualColumnNode
(
sourceResultSet,
sourceResultColumnList.elementAt(index),
index + 1,
getContextManager()
)
);
/* Mark the ResultColumn as being referenced */
if (markReferenced)
{
resultColumn.setReferenced();
}
}
}
/**
* Walk the list and adjust the virtualColumnIds in the ResultColumns
* by the specified amount. If ResultColumn.expression is a VirtualColumnNode,
* then we adjust the columnId there as well.
*
* @param adjust The size of the increment.
*/
void adjustVirtualColumnIds(int adjust)
{
for (ResultColumn resultColumn : this)
{
resultColumn.adjustVirtualColumnId(adjust);
if (SanityManager.DEBUG)
{
if ( !
(resultColumn.getExpression() instanceof VirtualColumnNode))
{
SanityManager.THROWASSERT(
"resultColumn.getExpression() is expected to be " +
"instanceof VirtualColumnNode" +
" not " +
resultColumn.getExpression().getClass().getName());
}
}
((VirtualColumnNode) resultColumn.getExpression()).columnId += adjust;
}
}
/**
* Project out any unreferenced ResultColumns from the list and
* reset the virtual column ids in the referenced ResultColumns.
* If all ResultColumns are projected out, then the list is not empty.
*
* @exception StandardException Thrown on error
*/
void doProjection() throws StandardException
{
int numDeleted = 0;
ResultColumnList deletedRCL = new ResultColumnList(getContextManager());
for (ResultColumn resultColumn : this)
{
/* RC's for FromBaseTables are marked as referenced during binding.
* For other nodes, namely JoinNodes, we need to go 1 level
* down the RC/VCN chain to see if the RC is referenced. This is
* because we propagate the referencing info from the bottom up.
*/
if ((! resultColumn.isReferenced()) &&
(resultColumn.getExpression() instanceof VirtualColumnNode) &&
!(((VirtualColumnNode) resultColumn.getExpression()).getSourceColumn().isReferenced()))
{
// Remember the RC to delete when done
deletedRCL.addElement(resultColumn);
/* Remember how many we have deleted and decrement the
* VirtualColumnIds for all nodes which appear after us
* in the list.
*/
numDeleted++;
}
else
{
/* Decrement the VirtualColumnId for each node in the list
* after the 1st deleted one.
*/
if (numDeleted >= 1)
resultColumn.adjustVirtualColumnId( - numDeleted);
/* Make sure that the RC is marked as referenced! */
resultColumn.setReferenced();
}
}
// Go back and delete the RCs to be delete from the list
for (int index = 0; index < deletedRCL.size(); index++)
{
removeElement(deletedRCL.elementAt(index));
}
}
/**
* Check the uniqueness of the column names within a column list.
*
* @param errForGenCols Raise an error for any generated column names.
*
* @return String The first duplicate column name, if any.
*/
String verifyUniqueNames(boolean errForGenCols)
throws StandardException
{
HashSet<String> seenNames = new HashSet<String>(size() + 2, 0.999f);
for (ResultColumn rc : this)
{
if (errForGenCols && rc.isNameGenerated())
throw StandardException.newException(SQLState.LANG_DB2_VIEW_REQUIRES_COLUMN_NAMES);
/* Verify that this column's name is unique within the list */
String colName = rc.getName();
boolean alreadySeen = !seenNames.add(colName);
if (alreadySeen)
{
return colName;
}
}
/* No duplicate column names */
return null;
}
/**
* Validate the derived column list (DCL) and propagate the info
* from the list to the final ResultColumnList.
*
* @param derivedRCL The derived column list
* @param tableName The table name for the FromTable
*
* @exception StandardException Thrown on error
*/
void propagateDCLInfo(ResultColumnList derivedRCL, String tableName)
throws StandardException
{
String duplicateColName;
/* Do both lists, if supplied by user, have the same degree? */
if (derivedRCL.size() != size() &&
! derivedRCL.getCountMismatchAllowed())
{
if (visibleSize() != derivedRCL.visibleSize()) {
throw StandardException.newException(SQLState.LANG_DERIVED_COLUMN_LIST_MISMATCH, tableName);
}
}
/* Check the uniqueness of the column names within the derived list */
duplicateColName = derivedRCL.verifyUniqueNames(false);
if (duplicateColName != null)
{
throw StandardException.newException(SQLState.LANG_DUPLICATE_COLUMN_NAME_DERIVED, duplicateColName);
}
/* We can finally copy the derived names into the final list */
copyResultColumnNames(derivedRCL);
}
/**
* Look for and reject ? parameters under ResultColumns. This is done for
* SELECT statements.
*
* @exception StandardException Thrown if a ? parameter found directly
* under a ResultColumn
*/
void rejectParameters() throws StandardException
{
for (ResultColumn rc : this)
{
rc.rejectParameter();
}
}
/**
* Check for (and reject) XML values directly under the ResultColumns.
* This is done for SELECT/VALUES statements. We reject values
* in this case because JDBC does not define an XML type/binding
* and thus there's no standard way to pass such a type back
* to a JDBC application.
*
* Note that we DO allow an XML column in a top-level RCL
* IF that column was added to the RCL by _us_ instead of
* by the user. For example, if we have a table:
*
* create table t1 (i int, x xml)
*
* and the user query is:
*
* select i from t1 order by x
*
* the "x" column will be added (internally) to the RCL
* as part of ORDER BY processing--and so we need to
* allow that XML column to be bound without throwing
* an error. If, as in this case, the XML column reference
* is invalid (we can't use ORDER BY on an XML column because
* XML values aren't ordered), a more appropriate error
* message should be returned to the user in later processing.
* If we didn't allow for this, the user would get an
* error saying that XML columns are not valid as part
* of the result set--but as far as s/he knows, there
* isn't such a column: only "i" is supposed to be returned
* (the RC for "x" was added to the RCL by _us_ as part of
* ORDER BY processing).
*
* ASSUMPTION: Any RCs that are generated internally and
* added to this RCL (before this RCL is bound) are added
* at the _end_ of the list. If that's true, then any
* RC with an index greater than the size of the initial
* (user-specified) list must have been added internally
* and will not be returned to the user.
*
* @exception StandardException Thrown if an XML value found
* directly under a ResultColumn
*/
void rejectXMLValues() throws StandardException
{
int sz = size();
for (int i = 1; i <= sz; i++) {
if (i > initialListSize)
// this RC was generated internally and will not
// be returned to the user, so don't throw error.
continue;
ResultColumn rc = getResultColumn(i);
if ((rc != null) && (rc.getType() != null) &&
rc.getType().getTypeId().isXMLTypeId())
{ // Disallow it.
throw StandardException.newException(
SQLState.LANG_ATTEMPT_TO_SELECT_XML);
}
}
}
/**
* Set the resultSetNumber in all of the ResultColumns.
*
* @param resultSetNumber The resultSetNumber
*/
void setResultSetNumber(int resultSetNumber)
{
for (ResultColumn rc : this)
{
rc.setResultSetNumber(resultSetNumber);
}
}
/**
* Mark all of the ResultColumns as redundant.
* Useful when chopping a ResultSetNode out of a tree when there are
* still references to its RCL.
*/
void setRedundant()
{
for (ResultColumn rc : this)
{
rc.setRedundant();
}
}
/**
* Verify that all of the columns in the SET clause of a positioned update
* appear in the cursor's FOR UPDATE OF list.
*
* @param cursorStmt the statement that owns the cursor
* @param cursorName The cursor's name.
*
* @exception StandardException Thrown on error
*/
void checkColumnUpdateability(
ExecPreparedStatement cursorStmt, String cursorName)
throws StandardException
{
for (ResultColumn resultColumn : this)
{
if (resultColumn.updated() &&
!cursorStmt.isUpdateColumn(resultColumn.getName()))
{
throw StandardException.newException(SQLState.LANG_COLUMN_NOT_UPDATABLE_IN_CURSOR,
resultColumn.getName(),
cursorName);
}
}
}
/**
* Set up the result expressions for a UNION, INTERSECT, or EXCEPT:
* o Verify union type compatiblity
* o Get dominant type for result (type + max length + nullability)
* o Create a new ColumnReference with dominant type and name of from this
* RCL and make that the new expression.
* o Set the type info for in the ResultColumn to the dominant type
*
* NOTE - We are assuming that caller has generated a new RCL for the UNION
* with the same names as the left side's RCL and copies of the expressions.
*
* @param otherRCL RCL from other side of the UNION.
* @param tableNumber The tableNumber for the UNION.
* @param level The nesting level for the UNION.
* @param operatorName "UNION", "INTERSECT", or "EXCEPT"
*
* @exception StandardException Thrown on error
*/
void setUnionResultExpression(ResultColumnList otherRCL,
int tableNumber,
int level,
String operatorName)
throws StandardException
{
TableName dummyTN;
if (SanityManager.DEBUG)
{
if (visibleSize() != otherRCL.visibleSize())
{
SanityManager.THROWASSERT(
"visibleSize() = (" +
visibleSize() +
") is expected to equal otherRCL.visibleSize (" +
otherRCL.visibleSize() +
")");
}
// Generated grouping columns and unselected ORDER BY columns
// should have been removed for the RCL of a SetOperatorNode, so
// that size and visible size are equal (DERBY-3764).
SanityManager.ASSERT(size() == visibleSize(),
"size() and visibleSize() should be equal");
}
/* Make a dummy TableName to be shared by all new CRs */
dummyTN = new TableName(null, null, getContextManager());
int size = visibleSize();
for (int index = 0; index < size; index++)
{
ColumnReference newCR;
ResultColumn thisRC = elementAt(index);
ResultColumn otherRC = otherRCL.elementAt(index);
ValueNode thisExpr = thisRC.getExpression();
ValueNode otherExpr = otherRC.getExpression();
// If there is one row that is not 'autoincrement', the Union should
// not be 'autoincrement'.
if (!otherRC.isAutoincrementGenerated() && thisRC.isAutoincrementGenerated())
{
thisRC.resetAutoincrementGenerated();
}
/*
** If there are ? parameters in the ResultColumnList of a row
** in a table constructor, their types will not be set. Just skip
** these - their types will be set later. Each ? parameter will
** get the type of the first non-? in its column, so it can't
** affect the final dominant type. It's possible that all the
** rows for a particular column will have ? parameters - this is
** an error condition that will be caught later.
*/
TypeId thisTypeId = thisExpr.getTypeId();
if (thisTypeId == null)
continue;
TypeId otherTypeId = otherExpr.getTypeId();
if (otherTypeId == null)
continue;
/*
** Check type compatability.
*/
ClassFactory cf = getClassFactory();
if ( !unionCompatible( thisExpr, otherExpr ) )
{
throw StandardException.newException(SQLState.LANG_NOT_UNION_COMPATIBLE,
thisTypeId.getSQLTypeName(),
otherTypeId.getSQLTypeName(),
operatorName);
}
DataTypeDescriptor resultType = thisExpr.getTypeServices().getDominantType(
otherExpr.getTypeServices(),
cf);
newCR = new ColumnReference(
thisRC.getName(), dummyTN, getContextManager());
newCR.setType(resultType);
/* Set the tableNumber and nesting levels in newCR.
* If thisExpr is not a CR, then newCR cannot be
* correlated, hence source and nesting levels are
* the same.
*/
if (thisExpr instanceof ColumnReference)
{
newCR.copyFields((ColumnReference) thisExpr);
}
else
{
newCR.setNestingLevel(level);
newCR.setSourceLevel(level);
}
newCR.setTableNumber(tableNumber);
thisRC.setExpression(newCR);
thisRC.setType(
thisRC.getTypeServices().getDominantType(
otherRC.getTypeServices(), cf));
/* DB2 requires both sides of union to have same name for the result to
* have that name. Otherwise, leave it or set it to a generated name */
if (thisRC.getName() != null && !thisRC.isNameGenerated() &&
otherRC.getName() != null)
{
/* Result name needs to be changed */
if (otherRC.isNameGenerated())
{
thisRC.setName(otherRC.getName());
thisRC.setNameGenerated(true);
}
else if (!thisRC.getName().equals(otherRC.getName()))
{
/* Both sides have user specified names that don't match */
thisRC.setName(null);
thisRC.guaranteeColumnName();
thisRC.setNameGenerated(true);
}
}
}
}
/**
* Return true if the types of two expressions are union compatible. The rules for union
* compatibility are found in the SQL Standard, part 2, section 7.3 (<query expression>),
* syntax rule 20.b.ii. That in turn, refers you to section 9.3 (Result of data type combinations).
* See, for instance, <a href="https://issues.apache.org/jira/browse/DERBY-4692">DERBY-4692</a>.
*
* This logic may enforce only a weaker set of rules. Here is the original comment
* on the original logic: "We want to make sure that the types are assignable in either direction
* and they are comparable." We may need to revisit this code to make it conform to the
* Standard.
*/
private boolean unionCompatible( ValueNode left, ValueNode right )
throws StandardException
{
TypeId leftTypeId = left.getTypeId();
TypeId rightTypeId = right.getTypeId();
ClassFactory cf = getClassFactory();
if (
!left.getTypeCompiler().storable(rightTypeId, cf) &&
!right.getTypeCompiler().storable(leftTypeId, cf)
)
{ return false; }
if ( leftTypeId.isBooleanTypeId() != rightTypeId.isBooleanTypeId() ) { return false; }
return true;
}
/**
* Do the 2 RCLs have the same type and length.
* This is useful for UNIONs when deciding whether a NormalizeResultSet is required.
*
* @param otherRCL The other RCL.
*
* @return boolean Whether or not there is an exact UNION type match on the 2 RCLs.
*/
boolean isExactTypeAndLengthMatch(ResultColumnList otherRCL)
throws StandardException
{
if (SanityManager.DEBUG) {
// The visible size of the two RCLs must be equal.
SanityManager.ASSERT(visibleSize() == otherRCL.visibleSize(),
"visibleSize() should match");
// The generated grouping columns and unselected ORDER BY columns
// should have been removed from the RCL of the SetOperatorNode,
// so size and visible size should be equal (DERBY-3764).
SanityManager.ASSERT(size() == visibleSize(),
"size() and visibleSize() should match");
}
int size = visibleSize();
for (int index = 0; index < size; index++)
{
ResultColumn thisRC = elementAt(index);
ResultColumn otherRC = otherRCL.elementAt(index);
if (! thisRC.getTypeServices().isExactTypeAndLengthMatch(
otherRC.getTypeServices() ))
{
return false;
}
}
return true;
}
/**
* Does the column list contain any of the given column positions
* that are updated? Implements same named routine in UpdateList.
*
* @param columns An array of column positions
*
* @return True if this column list contains any of the given columns
*/
public boolean updateOverlaps(int[] columns)
{
for (ResultColumn rc : this)
{
if ( ! rc.updated())
continue;
int column = rc.getColumnPosition();
for (int i = 0; i < columns.length; i++)
{
if (columns[i] == column)
return true;
}
}
return false;
}
/**
* Return an array that contains references to the columns in this list
* sorted by position.
*
* @return The sorted array.
*/
ResultColumn[] getSortedByPosition()
{
int size = size();
ResultColumn[] result;
/*
** Form an array of the original ResultColumns
*/
result = new ResultColumn[size];
/*
** Put the ResultColumns in the array
*/
for (int index = 0; index < size; index++)
{
result[index] = elementAt(index);
}
/*
** Sort the array by column position
*/
java.util.Arrays.sort(result);
return result;
}
/**
* Return an array of all my column positions, sorted in
* ascending order.
*
* @return a sorted array
*/
public int[] sortMe()
{
ResultColumn[] sortedResultColumns = getSortedByPosition();
int[] sortedColumnIds = new int[sortedResultColumns.length];
for (int ix = 0; ix < sortedResultColumns.length; ix++)
{
sortedColumnIds[ix] = sortedResultColumns[ix].getColumnPosition();
}
return sortedColumnIds;
}
/**
* Expand this ResultColumnList by adding all columns from the given
* table that are not in this list. The result is sorted by column
* position.
*
* @param td The TableDescriptor for the table in question
* @param tableName The name of the table as given in the query
*
* @return A new ResultColumnList expanded to include all columns in
* the given table.
*
* @exception StandardException Thrown on error
*/
ResultColumnList expandToAll(TableDescriptor td,
TableName tableName)
throws StandardException
{
ResultColumn rc;
ColumnDescriptor cd;
ResultColumnList retval;
ResultColumn[] originalRCS;
int posn;
/* Get a new ResultColumnList */
retval = new ResultColumnList(getContextManager());
/*
** Form a sorted array of the ResultColumns
*/
originalRCS = getSortedByPosition();
posn = 0;
/* Iterate through the ColumnDescriptors for the given table */
ColumnDescriptorList cdl = td.getColumnDescriptorList();
int cdlSize = cdl.size();
for (int index = 0; index < cdlSize; index++)
{
cd = cdl.elementAt(index);
if ((posn < originalRCS.length) &&
(cd.getPosition() == originalRCS[posn].getColumnPosition()))
{
rc = originalRCS[posn];
posn++;
}
else
{
/* Build a ResultColumn/ColumnReference pair for the column */
rc = makeColumnReferenceFromName( tableName, cd.getColumnName() );
/* Bind the new ResultColumn */
rc.bindResultColumnByPosition(td, cd.getPosition());
}
/* Add the ResultColumn to the list */
retval.addResultColumn(rc);
}
if (SanityManager.DEBUG)
SanityManager.ASSERT(posn == originalRCS.length,
"ResultColumns in original list not added to expanded ResultColumnList");
return retval;
}
/**
* Bind any untyped null nodes to the types in the given ResultColumnList.
* Nodes that don't know their type may pass down nulls to
* children nodes. In the case of something like a union, it knows
* to try its right and left result sets against each other.
* But if a null reaches us, it means we have a null type that
* we don't know how to handle.
*
* @param bindingRCL The ResultColumnList with the types to bind to.
*
* @exception StandardException Thrown on error
*/
void bindUntypedNullsToResultColumns(ResultColumnList bindingRCL)
throws StandardException
{
if (bindingRCL == null)
{
throw StandardException.newException(SQLState.LANG_NULL_IN_VALUES_CLAUSE);
}
if (SanityManager.DEBUG)
SanityManager.ASSERT(bindingRCL.size() >= this.size(),
"More columns in result column list than in base table");
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn bindingRC = bindingRCL.elementAt(index);
ResultColumn thisRC = elementAt(index);
thisRC.typeUntypedNullExpression(bindingRC);
}
}
/**
* Mark all the columns in this list as updated by an update statement.
*/
void markUpdated()
{
for (ResultColumn rc : this)
{
rc.markUpdated();
}
}
/**
* Mark all the (base) columns in this list as updatable by a positioned update
* statement. This is necessary
* for positioned update statements, because we expand the column list
* to include all the columns in the base table, and we need to be able
* to tell which ones the user is really trying to update so we can
* determine correctly whether all the updated columns are in the
* "for update" list.
*/
void markUpdatableByCursor()
{
for (ResultColumn rc : this)
{
// Determine whether the column is a base column and
// not a derived column, and, additionally,
// verify that the column was not aliased.
//
if (rc.getSourceTableName() != null &&
rc.getExpression() != null &&
rc.getExpression().getColumnName().equals(rc.getName()))
rc.markUpdatableByCursor();
}
}
/**
* Verify that all of the column names in this list are contained
* within the ColumnDefinitionNodes within the TableElementList.
*
*
* @return String The 1st column name, if any, that is not in the list.
*/
String verifyCreateConstraintColumnList(TableElementList tel)
{
for (ResultColumn rc : this)
{
String colName = rc.getName();
if (! tel.containsColumnName(colName))
{
return colName;
}
}
return null;
}
/**
* Export the result column names to the passed in String[].
*
* @param columnNames String[] to hold the column names.
*/
void exportNames(String[] columnNames)
{
if (SanityManager.DEBUG)
{
if (size() != columnNames.length)
{
SanityManager.THROWASSERT(
"size() (" +
size() +
") is expected to equal columnNames.length (" +
columnNames.length +
")");
}
}
int size = size();
for (int index = 0; index < size; index++)
{
columnNames[index] = elementAt(index).getName();
}
}
/**
* Given a ResultColumn at the next deepest level in the tree,
* search this RCL for its parent ResultColumn.
*
* @param childRC The child ResultColumn
*
* @return ResultColumn The parent ResultColumn
*/
ResultColumn findParentResultColumn(ResultColumn childRC)
{
ResultColumn parentRC = null;
for (ResultColumn rc : this)
{
if (rc.getExpression() instanceof ColumnReference)
{
ColumnReference cr = (ColumnReference) rc.getExpression();
if (cr.getSource() == childRC)
{
parentRC = rc;
break;
}
}
else if (rc.getExpression() instanceof VirtualColumnNode)
{
VirtualColumnNode vcn = (VirtualColumnNode) rc.getExpression();
if (vcn.getSourceColumn() == childRC)
{
parentRC = rc;
break;
}
}
}
return parentRC;
}
/**
* Mark as updatable all the columns in this result column list
* that match the columns in the given update column list.
*
* @param updateColumns A ResultColumnList representing the columns
* to be updated.
*/
void markUpdated(ResultColumnList updateColumns)
{
for (ResultColumn updateColumn : updateColumns)
{
ResultColumn resultColumn = getResultColumn(updateColumn.getName());
/*
** This ResultColumnList may not be bound yet - for update
** statements, we mark the updated columns *before* we bind
** the RCL. This ordering is important because we add columns
** to the RCL after marking the update columns and before
** binding.
**
** So, it can happen that there is an invalid column name in
** the list. This condition will cause an exception when the
** RCL is bound. Just ignore it for now.
*/
if (resultColumn != null)
{
resultColumn.markUpdated();
}
}
}
/**
* Mark all the columns in the select sql that this result column list represents
* as updatable if they match the columns in the given update column list.
*
* @param updateColumns A list representing the columns
* to be updated.
*/
void markColumnsInSelectListUpdatableByCursor(List<String> updateColumns)
{
commonCodeForUpdatableByCursor(updateColumns, true);
}
/**
* dealingWithSelectResultColumnList true means we are dealing with
* ResultColumnList for a select sql. When dealing with ResultColumnList for
* select sql, it is possible that not all the updatable columns are
* projected in the select column list and hence it is possible that we may
* not find the column to be updated in the ResultColumnList and that is why
* special handling is required when dealingWithSelectResultColumnList is true.
* eg select c11, c13 from t1 for update of c11, c12
* In the eg above, we will find updatable column c11 in the select column
* list but we will not find updatable column c12 in the select column list
*/
private void commonCodeForUpdatableByCursor(
List<String> updateColumns,
boolean dealingWithSelectResultColumnList)
{
/*
** If there is no update column list, or the list is empty, then it means that
** all the columns which have a base table associated with them are updatable.
*/
if ( (updateColumns == null) || (updateColumns.isEmpty()) )
{
markUpdatableByCursor();
}
else
{
int ucSize = updateColumns.size();
ResultColumn resultColumn;
String columnName;
for (int index = 0; index < ucSize; index++)
{
columnName = updateColumns.get(index);
resultColumn = getResultColumn(columnName);
if (SanityManager.DEBUG)
{
if (resultColumn == null && !dealingWithSelectResultColumnList)
{
SanityManager.THROWASSERT("No result column found with name " +
columnName);
}
}
//Following if means the column specified in FOR UPDATE clause is not
//part of the select list
if (resultColumn == null && dealingWithSelectResultColumnList)
continue;
resultColumn.markUpdatableByCursor();
}
}
}
/**
* Mark as updatable all the columns in this result column list
* that match the columns in the given update column list
*
* @param updateColumns A list representing the columns
* to be updated.
*/
void markUpdatableByCursor(List<String> updateColumns)
{
commonCodeForUpdatableByCursor(updateColumns, false);
}
/**
* Returns true if the given column position is for a column that will
* be or could be updated by the positioned update of a cursor.
*
* @param columnPosition The position of the column in question
*
* @return true if the column is updatable
*/
boolean updatableByCursor(int columnPosition)
{
return getResultColumn(columnPosition).updatableByCursor();
}
/**
* Return whether or not this RCL can be flattened out of a tree.
* It can only be flattened if the expressions are all cloneable.
*
* @return boolean Whether or not this RCL can be flattened out of a tree.
*/
boolean isCloneable()
{
boolean retcode = true;
for (ResultColumn rc : this)
{
if (! rc.getExpression().isCloneable())
{
retcode = false;
break;
}
}
return retcode;
}
/**
* Remap all ColumnReferences in this tree to be clones of the
* underlying expression.
*
* @exception StandardException Thrown on error
*/
void remapColumnReferencesToExpressions() throws StandardException
{
for (ResultColumn rc : this)
{
// The expression may be null if this column is an identity
// column generated always. If the expression is not null, it
// is a ColumnReference; we call through to the ColumnReference
// to give it a chance to remap itself from the outer query
// node to this one.
if (rc.getExpression() != null)
rc.setExpression(
rc.getExpression().remapColumnReferencesToExpressions());
}
}
/*
** Indicate that the conglomerate is an index, so we need to generate a
** RowLocation as the last column of the result set.
**
** @param cid The conglomerate id of the index
*/
void setIndexRow(long cid, boolean forUpdate)
{
indexRow = true;
conglomerateId = cid;
this.forUpdate = forUpdate;
}
/* Debugging methods */
/**
* Verify that all ResultColumns and their expressions have type information
* and that the type information between the respective RCs and
* expressions matches.
*
* @return boolean Whether or not the type information is consistent
*/
boolean hasConsistentTypeInfo() throws StandardException
{
boolean isConsistent = true;
if (SanityManager.DEBUG)
{
for (ResultColumn rc : this)
{
ValueNode expr = rc.getExpression();
DataTypeDescriptor rcDTS = rc.getTypeServices();
DataTypeDescriptor exDTS = expr.getTypeServices();
if (rcDTS == null || exDTS == null)
{
isConsistent = false;
break;
}
if (rcDTS.getClass().getName() !=
exDTS.getClass().getName())
{
isConsistent = false;
break;
}
}
}
return isConsistent;
}
/**
* Return whether or not this RCL contains an AllResultColumn.
* This is useful when dealing with SELECT * views which
* reference tables that may have had columns added to them via
* ALTER TABLE since the view was created.
*
* @return Whether or not this RCL contains an AllResultColumn.
*/
boolean containsAllResultColumn()
{
boolean containsAllResultColumn = false;
int size = size();
for (int index = 0; index < size; index++)
{
if (elementAt(index) instanceof AllResultColumn)
{
containsAllResultColumn = true;
break;
}
}
return containsAllResultColumn;
}
/**
* Count the number of RCs in the list that are referenced.
*
* @return The number of RCs in the list that are referenced.
*/
int countReferencedColumns()
{
int numReferenced = 0;
for (ResultColumn rc : this)
{
if (rc.isReferenced())
{
numReferenced++;
}
}
return numReferenced;
}
/**
* Record the column ids of the referenced columns in the specified array.
*
* @param idArray int[] for column ids
* @param basis 0 (for 0-based ids) or 1 (for 1-based ids)
*/
void recordColumnReferences(int[] idArray, int basis)
{
int currArrayElement = 0;
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn rc = elementAt(index);
if (rc.isReferenced())
{
idArray[currArrayElement++] = index + basis;
}
}
}
/**
* Get the position of first result column with the given name.
*
* @param name Name of the column
* @param basis 0 (for 0-based ids) or 1 (for 1-based ids)
*/
int getPosition( String name, int basis )
{
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn rc = elementAt(index);
if ( name.equals( rc.getName() ) )
{
return index + basis;
}
}
return -1;
}
/**
* Record the top level ColumnReferences in the specified array
* and table map
* This is useful when checking for uniqueness conditions.
* NOTE: All top level CRs assumed to be from the same table.
* The size of the array is expected to be the # of columns
* in the table of interest + 1, so we use 1-base column #s.
*
* @param colArray1 boolean[] for columns
* @param tableColMap JBitSet[] for tables
* @param tableNumber Table number of column references
*/
void recordColumnReferences(boolean[] colArray1, JBitSet[] tableColMap,
int tableNumber)
{
for (ResultColumn rc : this)
{
int columnNumber;
if (! (rc.getExpression() instanceof ColumnReference))
{
continue;
}
columnNumber = ((ColumnReference) rc.getExpression()).getColumnNumber();
colArray1[columnNumber] = true;
tableColMap[tableNumber].set(columnNumber);
}
}
/**
* Return whether or not all of the RCs in the list whose
* expressions are ColumnReferences are
* from the same table. One place this
* is useful for distinct elimination based on the existence
* of a uniqueness condition.
*
* @return -1 if all of the top level CRs in the RCL
* are not ColumnReferences from the same table,
* else the tableNumber
*/
int allTopCRsFromSameTable()
{
int tableNumber = -1;
for (ResultColumn rc : this)
{
ValueNode vn = rc.getExpression();
if (! (vn instanceof ColumnReference))
{
continue;
}
// Remember the tableNumber from the first CR
ColumnReference cr = (ColumnReference) vn;
if (tableNumber == -1)
{
tableNumber = cr.getTableNumber();
}
else if (tableNumber != cr.getTableNumber())
{
return -1;
}
}
return tableNumber;
}
/**
* Clear the column references from the RCL. (Restore RCL back to a state
* where none of the RCs are marked as referenced.)
*/
void clearColumnReferences()
{
for (ResultColumn rc : this)
{
if (rc.isReferenced())
{
rc.setUnreferenced();
}
}
}
/**
* Copy the referenced RCs from this list to the supplied target list.
*
* @param targetList The list to copy to
*/
void copyReferencedColumnsToNewList(ResultColumnList targetList)
{
for (ResultColumn rc : this)
{
if (rc.isReferenced())
{
targetList.addElement(rc);
}
}
}
/**
* Copy the RCs from this list to the supplied target list.
*
* @param targetList The list to copy to,
* @param copyList 1 based bitMap we copy columns associated with set bits.
*/
void copyColumnsToNewList(
ResultColumnList targetList, FormatableBitSet copyList)
{
for (ResultColumn rc : this)
{
if (copyList.isSet(rc.getColumnPosition()))
{
targetList.addElement(rc);
}
}
}
/**
* Get a FormatableBitSet of the columns referenced in this rcl
*
* @return the FormatableBitSet
*/
FormatableBitSet getColumnReferenceMap()
{
FormatableBitSet colMap = new FormatableBitSet(size());
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn rc = elementAt(index);
if (rc.isReferenced())
{
colMap.set(index);
}
}
return colMap;
}
/**
* Or in any isReferenced booleans from the virtual column chain. That is the isReferenced bits on each
* ResultColumn on the list will be set if the ResultColumn is referenced or if any VirtualColumnNode in its
* expression chain refers to a referenced column.
*/
void pullVirtualIsReferenced()
{
for (ResultColumn rc : this)
{
rc.pullVirtualIsReferenced();
}
} // end of pullVirtualIsReferenced
void clearTableNames()
{
for (ResultColumn rc : this)
{
rc.clearTableName();
}
}
/**
* Set the value of whether or not a count mismatch is allowed between
* this RCL, as a derived column list, and an underlying RCL. This is allowed
* for SELECT * views when an underlying table has had columns added to it
* via ALTER TABLE.
*
* @param allowed Whether or not a mismatch is allowed.
*/
protected void setCountMismatchAllowed(boolean allowed)
{
countMismatchAllowed = allowed;
}
/**
* Return whether or not a count mismatch is allowed between this RCL,
* as a derived column list, and an underlying RCL. This is allowed
* for SELECT * views when an underlying table has had columns added to it
* via ALTER TABLE.
*
* return Whether or not a mismatch is allowed.
*/
protected boolean getCountMismatchAllowed()
{
return countMismatchAllowed;
}
/**
* Get the size of all the columns added
* together. Does <B>NOT</B> include the
* column overhead that the store requires.
* Also, will be a very rough estimate for
* user types.
*
* @return the size
*/
int getTotalColumnSize()
{
int colSize = 0;
for (ResultColumn rc : this)
{
colSize += rc.getMaximumColumnSize();
}
return colSize;
}
/**
* Generate an RCL to match the contents of a ResultSetMetaData.
* This is useful when dealing with VTIs.
*
* @param rsmd The ResultSetMetaData.
* @param tableName The TableName for the BCNs.
* @param javaClassName The name of the VTI
*
* @exception StandardException Thrown on error
*/
void createListFromResultSetMetaData(ResultSetMetaData rsmd,
TableName tableName,
String javaClassName)
throws StandardException
{
try
{
// JDBC columns #s are 1-based
// Check to make sure # of columns >= 1
int numColumns = rsmd.getColumnCount();
if (numColumns <= 0)
{
throw StandardException.newException(SQLState.LANG_INVALID_V_T_I_COLUMN_COUNT,
javaClassName, String.valueOf(numColumns));
}
for (int index = 1; index <= numColumns; index++)
{
boolean nullableResult =
(rsmd.isNullable(index) != ResultSetMetaData.columnNoNulls);
TypeId cti;
int jdbcColumnType = rsmd.getColumnType(index);
switch (jdbcColumnType) {
case Types.JAVA_OBJECT:
case Types.OTHER:
{
cti = TypeId.getUserDefinedTypeId(
rsmd.getColumnTypeName(index));
break;
}
default:
{
cti = TypeId.getBuiltInTypeId(jdbcColumnType);
break;
}
}
// Handle the case where a VTI returns a bad column type
if (cti == null)
{
throw StandardException.newException(SQLState.LANG_BAD_J_D_B_C_TYPE_INFO, Integer.toString(index));
}
// Get the maximum byte storage for this column
int maxWidth;
/* Get maximum byte storage from rsmd for variable
* width types, set it to MAXINT for the long types,
* otherwise get it from the TypeId
*/
if (cti.variableLength())
{
maxWidth = rsmd.getColumnDisplaySize(index);
}
else if (jdbcColumnType == Types.LONGVARCHAR ||
jdbcColumnType == Types.LONGVARBINARY)
{
maxWidth = Integer.MAX_VALUE;
}
else
{
maxWidth = 0;
}
int precision = cti.isDecimalTypeId() ? rsmd.getPrecision(index) : 0;
int scale = cti.isDecimalTypeId() ? rsmd.getScale(index) : 0;
DataTypeDescriptor dts = new DataTypeDescriptor(cti,
precision,
scale,
nullableResult,
maxWidth);
addColumn( tableName, rsmd.getColumnName(index), dts );
}
}
catch (Throwable t)
{
if (t instanceof StandardException)
{
throw (StandardException) t;
}
else
{
throw StandardException.unexpectedUserException(t);
}
}
}
/**
* Add a column to the list given a table name, column name, and data type.
* Return the just-added column.
*
*/
public ResultColumn addColumn( TableName tableName, String columnName, DataTypeDescriptor dts )
throws StandardException
{
ValueNode bcn = new BaseColumnNode(columnName,
tableName,
dts,
getContextManager());
ResultColumn rc =
new ResultColumn(columnName, bcn, getContextManager());
rc.setType(dts);
addResultColumn(rc);
return rc;
}
/**
* Add an RC to the end of the list for the RID from an index.
* NOTE: RC.expression is a CurrentRowLocationNode. This was previously only used
* for non-select DML. We test for this node when generating the holder above
* and generate the expected code. (We really should create yet another new node
* type with its own code generation.)
*
* @exception StandardException Thrown on error
*/
void addRCForRID()
throws StandardException
{
ResultColumn rowLocationColumn;
CurrentRowLocationNode rowLocationNode;
/* Generate the RowLocation column */
rowLocationNode = new CurrentRowLocationNode(getContextManager());
rowLocationColumn =
new ResultColumn("", rowLocationNode, getContextManager());
rowLocationColumn.markGenerated();
/* Append to the ResultColumnList */
addResultColumn(rowLocationColumn);
}
/**
* Walk the list and mark all RCs as unreferenced. This is useful
* when recalculating which RCs are referenced at what level like
* when deciding which columns need to be returned from a non-matching
* index scan (as opposed to those returned from the base table).
*
* @exception StandardException Thrown on error
*/
void markAllUnreferenced()
throws StandardException
{
for (ResultColumn rc : this)
{
rc.setUnreferenced();
}
}
/**
* Determine if all of the RC.expressions are columns in the source result set.
* This is useful for determining if we need to do reflection
* at execution time.
*
* @param sourceRS The source ResultSet.
*
* @return Whether or not all of the RC.expressions are columns in the source result set.
*/
boolean allExpressionsAreColumns(ResultSetNode sourceRS)
{
for (ResultColumn rc : this)
{
//DERBY-4631
//Following if condition if true means that the
// ResultColumn is a join column for a RIGHT OUTER
// JOIN with USING/NATURAL clause. At execution
// time, a join column's value should be determined
// by generated code which is equivalent to
// COALESCE(leftTableJoinColumn,rightTableJoinColumn).
// By returning false here, we allow Derby to generate
// code for functionality equivalent to COALESCE to
// determine join column's value.
if (rc.isRightOuterJoinUsingClause())
return false;
ValueNode expr = rc.getExpression();
if (! (expr instanceof VirtualColumnNode) &&
! (expr instanceof ColumnReference))
{
return false;
}
/* If the expression is a VirtualColumnNode, make sure that the column
* is coming from the source result set, ie, that it is not a correlated
* column.
*/
if (expr instanceof VirtualColumnNode)
{
VirtualColumnNode vcn = (VirtualColumnNode) expr;
if (vcn.getSourceResultSet() != sourceRS)
{
vcn.setCorrelated();
return false;
}
}
/* Make sure this is not a correlated CR */
if (expr instanceof ColumnReference)
{
ColumnReference cr = (ColumnReference) expr;
if (cr.getCorrelated())
{
return false;
}
}
}
return true;
}
/**
* Map the source columns to these columns. Build an array to represent the mapping.
* For each RC, if the expression is simply a VCN or a CR then set the array element to be
* the virtual column number of the source RC. Otherwise, set the array element to
* -1.
* This is useful for determining if we need to do reflection
* at execution time.
* <p/>
* Also build an array of boolean for columns that point to the same virtual
* column and have types that are streamable to be able to determine if
* cloning is needed at execution time.
*
* @return Array representiong mapping of RCs to source RCs.
*/
ColumnMapping mapSourceColumns()
{
int[] mapArray = new int[size()];
boolean[] cloneMap = new boolean[size()];
ResultColumn resultColumn;
// key: virtual column number, value: index
Map<Integer,Integer> seenMap = new HashMap<Integer,Integer>();
int size = size();
for (int index = 0; index < size; index++)
{
resultColumn = elementAt(index);
if (resultColumn.getExpression() instanceof VirtualColumnNode)
{
VirtualColumnNode vcn = (VirtualColumnNode) resultColumn.getExpression();
// Can't deal with correlated VCNs
if (vcn.getCorrelated())
{
mapArray[index] = -1;
}
else
{
// Virtual column #s are 1-based
ResultColumn rc = vcn.getSourceColumn();
updateArrays(mapArray, cloneMap, seenMap, rc, index);
}
}
else if (resultColumn.isRightOuterJoinUsingClause())
{
mapArray[index] = -1;
}
else if (resultColumn.getExpression() instanceof ColumnReference)
{
ColumnReference cr = (ColumnReference) resultColumn.getExpression();
// Can't deal with correlated CRs
if (cr.getCorrelated())
{
mapArray[index] = -1;
}
else
{
// Virtual column #s are 1-based
ResultColumn rc = cr.getSource();
updateArrays(mapArray, cloneMap, seenMap, rc, index);
}
}
else
{
mapArray[index] = -1;
}
}
ColumnMapping result = new ColumnMapping(mapArray, cloneMap);
return result;
}
/** Set the nullability of every ResultColumn in this list
* @throws StandardException */
void setNullability(boolean nullability) throws StandardException
{
for (ResultColumn rc : this)
{
rc.setNullability(nullability);
}
}
/**
* Generate a FormatableBitSet representing the columns that are referenced in this RCL.
* The caller decides if they want this FormatableBitSet if every RC is referenced.
*
* @param positionedUpdate Whether or not the scan that the RCL
* belongs to is for update w/o a column list
* @param always Whether or not caller always wants a non-null FormatableBitSet if
* all RCs are referenced.
* @param onlyBCNs If true, only set bit if expression is a BaseColumnNode,
* otherwise set bit for all referenced RCs.
*
* @return The FormatableBitSet representing the referenced RCs.
*/
FormatableBitSet getReferencedFormatableBitSet(boolean positionedUpdate, boolean always, boolean onlyBCNs)
{
int index;
int colsAdded = 0;
int size = size();
FormatableBitSet newReferencedCols = new FormatableBitSet(size);
/*
** For an updatable cursor, we need
** all columns.
*/
if (positionedUpdate)
{
if (always)
{
/* Set all bits in the bit map */
for (index = 0; index < size; index++)
{
newReferencedCols.set(index);
}
return newReferencedCols;
}
else
{
return null;
}
}
for (index = 0; index < size; index++)
{
ResultColumn oldCol = elementAt(index);
if (oldCol.isReferenced())
{
/* Skip RCs whose expression is not a BCN
* when requested to do so.
*/
if (onlyBCNs && ! (oldCol.getExpression() instanceof BaseColumnNode))
{
boolean skipable =
( ! (oldCol.getExpression() instanceof BaseColumnNode) ) &&
( ! (oldCol.getExpression() instanceof CurrentRowLocationNode) );
if ( skipable )
{
continue;
}
}
newReferencedCols.set(index);
colsAdded++;
}
}
/* Return the FormatableBitSet if not all RCs are referenced or if
* the caller always wants the FormatableBitSet returned.
*/
if (colsAdded != index || always)
{
return newReferencedCols;
}
else
{
return null;
}
}
/**
* Create a new, compacted RCL based on the referenced RCs
* in this list. If the RCL being compacted is for an
* updatable scan, then we simply return this.
*
* The caller tells us whether or not they want a new list
* if there is no compaction because all RCs are referenced.
* This is useful in the case where the caller needs a new
* RCL for existing RCs so that it can augment the new list.
*
* @param positionedUpdate Whether or not the scan that the RCL
* belongs to is for update w/o a column list
* @param always Whether or not caller always wants a new RCL
*
* @return The compacted RCL if compaction occurred, otherwise return this RCL.
*
* @exception StandardException Thrown on error
*/
ResultColumnList compactColumns(boolean positionedUpdate, boolean always)
throws StandardException
{
int index;
int colsAdded = 0;
/*
** For an updatable cursor, we need
** all columns.
*/
if (positionedUpdate)
{
return this;
}
ResultColumnList newCols = new ResultColumnList(getContextManager());
int size = size();
for (index = 0; index < size; index++)
{
ResultColumn oldCol = elementAt(index);
if (oldCol.isReferenced())
{
newCols.addResultColumn(oldCol);
colsAdded++;
}
}
/* Return new RCL if we found unreferenced columns or if
* the caller always wants a new list.
*/
if (colsAdded != index || always)
{
return newCols;
}
else
{
return this;
}
}
/**
* Remove the columns which are join columns (in the
* joinColumns RCL) from this list. This is useful
* for a JOIN with a USING clause.
*
* @param joinColumns The list of join columns
*/
void removeJoinColumns(ResultColumnList joinColumns)
{
for (ResultColumn joinRC : joinColumns)
{
String columnName = joinRC.getName();
// columnName should always be non-null
if (SanityManager.DEBUG)
{
SanityManager.ASSERT(columnName != null,
"columnName should be non-null");
}
ResultColumn rightRC = getResultColumn(columnName);
// Remove the RC from this list.
if (rightRC != null)
{
removeElement(rightRC);
}
}
}
/**
* Get the join columns from this list.
* This is useful for a join with a USING clause.
* (ANSI specifies that the join columns appear 1st.)
*
* @param joinColumns A list of the join columns.
*
* @return A list of the join columns from this list
*/
ResultColumnList getJoinColumns(ResultColumnList joinColumns)
throws StandardException
{
ResultColumnList newRCL = new ResultColumnList(getContextManager());
/* Find all of the join columns and put them 1st on the
* new RCL.
*/
for (ResultColumn joinRC : joinColumns)
{
String columnName = joinRC.getName();
// columnName should always be non-null
if (SanityManager.DEBUG)
{
SanityManager.ASSERT(columnName != null,
"columnName should be non-null");
}
ResultColumn xferRC = getResultColumn(columnName);
if (xferRC == null) {
throw StandardException.newException(
SQLState.LANG_COLUMN_NOT_FOUND, columnName);
}
// Add the RC to the new list.
newRCL.addElement(xferRC);
}
return newRCL;
}
/**
* Reset the virtual column ids for all of the
* underlying RCs. (Virtual column ids are 1-based.)
*/
void resetVirtualColumnIds()
{
int size = size();
for (int index = 0; index < size; index++)
{
/* ResultColumns are 1-based */
elementAt(index).setVirtualColumnId(index + 1);
}
}
/**
* Return whether or not the same result row can be used for all
* rows returned by the associated ResultSet. This is possible
* if all entries in the list are constants or AggregateNodes.
*
* @return Whether or not the same result row can be used for all
* rows returned by the associated ResultSet.
*/
boolean reusableResult()
{
for (ResultColumn rc : this)
{
if ((rc.getExpression() instanceof ConstantNode) ||
(rc.getExpression() instanceof AggregateNode))
{
continue;
}
return false;
}
return true;
}
/**
* Get an array of column positions (1-based) for all the columns
* in this RCL. Assumes that all the columns are in the passed-in
* table
*
* @return the array of strings
*
* @exception throws StandardException on error
*/
int[] getColumnPositions( TableDescriptor td )
throws StandardException
{
int size = size();
int[] myPositions = new int[ size ];
String columnName;
ColumnDescriptor cd;
for ( int index = 0; index < size; index++ )
{
ResultColumn resultColumn = elementAt( index );
columnName = resultColumn.getName();
cd = td.getColumnDescriptor( columnName );
if ( cd == null )
{
throw StandardException.newException
( SQLState.LANG_COLUMN_NOT_FOUND_IN_TABLE, columnName, td.getQualifiedName() );
}
myPositions[ index ] = cd.getPosition();
}
return myPositions;
}
/**
* Get an array of strings for all the columns
* in this RCL.
*
* @return the array of strings
*/
String[] getColumnNames()
{
String strings[] = new String[size()];
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn resultColumn = elementAt(index);
strings[index] = resultColumn.getName();
}
return strings;
}
/**
* Replace any DEFAULTs with the associated tree for the default if
* allowed, or flag.
*
* @param ttd The TableDescriptor for the target table.
* @param tcl The RCL for the target table.
* @param allowDefaults true if allowed
*
* @exception StandardException Thrown on error
*/
void replaceOrForbidDefaults(TableDescriptor ttd,
ResultColumnList tcl,
boolean allowDefaults)
throws StandardException
{
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn rc = elementAt(index);
if (rc.isDefaultColumn())
{
if (!allowDefaults) {
throw StandardException.newException(
SQLState.LANG_INVALID_USE_OF_DEFAULT);
}
// DefaultNode defaultNode = (DefaultNode) rc.getExpression();
// Get ColumnDescriptor by name or by position?
ColumnDescriptor cd = null;
if (tcl == null)
{
cd = ttd.getColumnDescriptor(index + 1);
}
else if (index < tcl.size())
{
ResultColumn trc = tcl.elementAt(index);
cd = ttd.getColumnDescriptor(trc.getName());
}
// Too many RCs if no ColumnDescriptor
if (cd == null)
{
throw StandardException.newException(SQLState.LANG_TOO_MANY_RESULT_COLUMNS,
ttd.getQualifiedName());
}
if (cd.isAutoincrement())
{
rc.setAutoincrementGenerated();
} // end of if ()
DefaultInfoImpl defaultInfo = (DefaultInfoImpl) cd.getDefaultInfo();
//
// For generated columns, we don't have enough context at this
// point to bind the generation clause (the default) and
// unfortunately that step occurs very soon after defaults are substituted in. The
// parsing and binding of generation clauses happens considerably
// later on. At this juncture, we can be patient and just plug
// in a NULL literal as a placeholder. For generated columns,
// the generation clause tree is plugged in in DMLModStatementNode.parseAndBindGenerationClauses().
//
if ( (defaultInfo != null) && !defaultInfo.isGeneratedColumn() )
{
setDefault( rc, cd, defaultInfo );
}
else
{
rc.setExpression(
new UntypedNullConstantNode(getContextManager()));
rc.setWasDefaultColumn( true );
}
rc.setDefaultColumn(false);
}
}
}
/** Set the default in a ResultColumn */
void setDefault( ResultColumn rc, ColumnDescriptor cd, DefaultInfoImpl defaultInfo )
throws StandardException
{
/* Query is dependent on the DefaultDescriptor */
DefaultDescriptor defaultDescriptor = cd.getDefaultDescriptor(getDataDictionary());
getCompilerContext().createDependency(defaultDescriptor);
rc.setExpression
(
DefaultNode.parseDefault
(
defaultInfo.getDefaultText(),
getLanguageConnectionContext(),
getCompilerContext()
)
);
}
/**
* Walk the RCL and check for DEFAULTs. DEFAULTs
* are invalid at the time that this method is called,
* so we throw an exception if found.
* NOTE: The grammar allows:
* VALUES DEFAULT;
*
* @exception StandardException Thrown on error
*/
void checkForInvalidDefaults()
throws StandardException
{
for (ResultColumn rc : this)
{
if (rc.isAutoincrementGenerated())
continue;
if (rc.isDefaultColumn())
{
throw StandardException.newException(SQLState.LANG_INVALID_USE_OF_DEFAULT);
}
}
}
/**
* Verify that all of the RCs in this list are comparable.
*
* @exception StandardException Thrown on error
*/
void verifyAllOrderable()
throws StandardException
{
for (ResultColumn rc : this)
{
rc.verifyOrderable();
}
}
/**
* Build this ResultColumnList from a table description and
* an array of column IDs.
*
* @param table describes the table
* @param columnIDs column positions in that table (1-based)
*
* @exception StandardException Thrown on error
*/
public void populate
(
TableDescriptor table,
int[] columnIDs
)
throws StandardException
{
if ( columnIDs == null ) { return; }
int count = columnIDs.length;
String columnName;
int columnPosition;
ResultColumn rc;
for ( int i = 0; i < count; i++ )
{
columnPosition = columnIDs[ i ];
columnName = table.getColumnDescriptor( columnPosition ).getColumnName();
rc = makeColumnFromName( columnName );
addResultColumn( rc );
}
}
private ResultColumn makeColumnFromName( String columnName )
throws StandardException
{
return new ResultColumn(columnName, null, getContextManager());
}
private ResultColumn makeColumnReferenceFromName
(
TableName tableName,
String columnName
)
throws StandardException
{
ContextManager cm = getContextManager();
ResultColumn rc = new ResultColumn(
columnName,
new ColumnReference(columnName, tableName, cm),
cm
);
return rc;
}
/**
* check if any autoincrement or generated columns exist in the result column list.
* called from insert or update where you cannot insert/update the value
* of a generated or autoincrement column.
*
* @exception StandardException If the column is an ai column
*/
void forbidOverrides(ResultColumnList sourceRSRCL)
throws StandardException
{
forbidOverrides( sourceRSRCL, false );
}
/**
* check if any autoincrement or generated columns exist in the result column list.
* called from insert or update where you cannot insert/update the value
* of a generated or autoincrement column.
*
* @exception StandardException If the column is an ai column
*/
void forbidOverrides(ResultColumnList sourceRSRCL, boolean defaultsWereReplaced )
throws StandardException
{
int size = size();
for (int index = 0; index < size; index++)
{
ResultColumn rc = elementAt(index);
ResultColumn sourceRC =
(sourceRSRCL == null) ? null : sourceRSRCL.elementAt(index);
ColumnDescriptor cd = rc.getTableColumnDescriptor();
if ( (cd != null) && cd.hasGenerationClause() )
{
if ( !defaultsWereReplaced && (sourceRC != null) && !sourceRC.hasGenerationClause() && !sourceRC.wasDefaultColumn() )
{
throw StandardException.newException(SQLState.LANG_CANT_OVERRIDE_GENERATION_CLAUSE, rc.getName());
}
if ( sourceRC != null ) { sourceRC.setColumnDescriptor(cd.getTableDescriptor(), cd); }
}
if ((cd != null) && (cd.isAutoincrement()))
{
if (
( (sourceRC != null) && (sourceRC.isAutoincrementGenerated()) ) ||
( cd.isAutoincAlways() && defaultsWereReplaced )
)
{
sourceRC.setColumnDescriptor(cd.getTableDescriptor(), cd);
}else{
if(cd.isAutoincAlways())
throw StandardException.newException(SQLState.LANG_AI_CANNOT_MODIFY_AI,
rc.getName());
}
}
}
}
void incOrderBySelect()
{
orderBySelect++;
}
private void decOrderBySelect()
{
orderBySelect--;
}
int getOrderBySelect()
{
return orderBySelect;
}
public void copyOrderBySelect( ResultColumnList src)
{
orderBySelect = src.orderBySelect;
}
/* ****
* Take note of the size of this RCL _before_ we start
* processing/binding it. This is so that, at bind time,
* we can tell if any columns in the RCL were added
* internally by us (i.e. they were not specified by the
* user and thus will not be returned to the user).
*/
protected void markInitialSize() {
initialListSize = size();
}
private int numGeneratedColumns()
{
int numGenerated = 0;
int sz = size();
boolean inVisibleRange = false;
for (int i = sz - 1; i >= 0; i--)
{
ResultColumn rc = elementAt(i);
if (rc.isGenerated())
{
if (SanityManager.DEBUG) {
// We expect the generated columns to always be at the end of the list.
if (inVisibleRange)
SanityManager.THROWASSERT("Encountered generated column in expected visible range at rcl[" + i +"]");
}
numGenerated++;
} else {
// We are counting down, so as soon as we see one visible column, the rest should be th same
inVisibleRange = true;
}
}
return numGenerated;
}
/**
* @return the number of generated columns in this RCL.
*/
int numGeneratedColumnsForGroupBy()
{
int numGenerated = 0;
int sz = size();
for (int i = sz - 1; i >= 0; i--) {
ResultColumn rc = elementAt(i);
if (rc.isGenerated() && rc.isGroupingColumn())
{
numGenerated++;
}
}
return numGenerated;
}
/**
* Remove any generated columns from this RCL.
*/
void removeGeneratedGroupingColumns()
{
int sz = size();
for (int i = sz - 1; i >= 0; i--)
{
ResultColumn rc = elementAt(i);
if (rc.isGenerated() && rc.isGroupingColumn())
{
removeElementAt(i);
}
}
}
/**
* @return the number of columns that will be visible during execution.
* During compilation we can add columns for a group by/order by but these
* to an RCL but these are projected out during query execution.
*/
int visibleSize()
{
return size() - orderBySelect - numGeneratedColumns();
}
/**
* Convert this object to a String. See comments in QueryTreeNode.java
* for how this should be done for tree printing.
*
* @return This object as a String
*/
@Override
public String toString()
{
if (SanityManager.DEBUG) {
return "indexRow: " + indexRow + "\n" +
"orderBySelect: " + orderBySelect + "\n" +
(indexRow ? "conglomerateId: " + conglomerateId + "\n"
: "") +
super.toString();
} else {
return "";
}
}
private static boolean streamableType(ResultColumn rc) {
DataTypeDescriptor dtd = rc.getType();
TypeId s = TypeId.getBuiltInTypeId(dtd.getTypeName());
if (s != null) {
return s.streamStorable();
} else {
return false;
}
}
private static void updateArrays(int[] mapArray,
boolean[] cloneMap,
Map<Integer,Integer> seenMap,
ResultColumn rc,
int index) {
int vcId = rc.getVirtualColumnId();
mapArray[index] = vcId;
if (streamableType(rc)) {
Integer seenIndex = seenMap.get(Integer.valueOf(vcId));
if (seenIndex != null) {
// We have already mapped this column at index
// seenIndex, so mark occurence 2..n for cloning.
cloneMap[index] = true;
} else {
seenMap.put(vcId, index);
}
}
}
static class ColumnMapping {
final int[] mapArray;
final boolean[] cloneMap;
private ColumnMapping(int[] mapArray, boolean[] cloneMap) {
this.mapArray = mapArray;
this.cloneMap = cloneMap;
}
}
}