| /* |
| * 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. |
| */ |
| |
| import type { SqlBase } from 'druid-query-toolkit'; |
| import { |
| SqlColumn, |
| SqlExpression, |
| SqlFunction, |
| SqlLiteral, |
| SqlQuery, |
| SqlSetStatement, |
| SqlStar, |
| } from 'druid-query-toolkit'; |
| |
| import type { RowColumn } from './general'; |
| import { offsetToRowColumn } from './general'; |
| |
| export function prettyPrintSql(b: SqlBase): string { |
| return b.prettyTrim(50).toString(); |
| } |
| |
| export function timeFormatToSql(timeFormat: string): SqlExpression | undefined { |
| switch (timeFormat) { |
| case 'auto': |
| return SqlExpression.parse('TIME_PARSE(TRIM(?))'); |
| |
| case 'iso': |
| return SqlExpression.parse('TIME_PARSE(?)'); |
| |
| case 'posix': |
| return SqlExpression.parse(`MILLIS_TO_TIMESTAMP((?) * 1000)`); |
| |
| case 'millis': |
| return SqlExpression.parse(`MILLIS_TO_TIMESTAMP(?)`); |
| |
| case 'micro': |
| return SqlExpression.parse(`MILLIS_TO_TIMESTAMP((?) / 1000)`); |
| |
| case 'nano': |
| return SqlExpression.parse(`MILLIS_TO_TIMESTAMP((?) / 1000000)`); |
| |
| default: |
| return SqlExpression.parse(`TIME_PARSE(?, '${timeFormat}')`); |
| } |
| } |
| |
| export function convertToGroupByExpression(ex: SqlExpression): SqlExpression | undefined { |
| const underlyingExpression = ex.getUnderlyingExpression(); |
| if (!(underlyingExpression instanceof SqlFunction)) return; |
| |
| const args = underlyingExpression.args?.values; |
| if (!args) return; |
| |
| const interestingArgs = args.filter( |
| arg => !(arg instanceof SqlLiteral || arg instanceof SqlStar), |
| ); |
| if (interestingArgs.length !== 1) return; |
| |
| const newEx = interestingArgs[0]; |
| if (newEx instanceof SqlColumn) return newEx; |
| |
| return newEx.as((ex.getOutputName() || 'grouped').replace(/^[a-z]+_/i, '')); |
| } |
| |
| function extractQueryPrefix(text: string): string { |
| let q = SqlQuery.parse(text); |
| |
| // The parser will parse a SELECT query with a partitionedByClause and clusteredByClause but that is not valid, remove them from the query |
| if (!q.getIngestTable() && (q.partitionedByClause || q.clusteredByClause)) { |
| q = q.changePartitionedByClause(undefined).changeClusteredByClause(undefined); |
| } |
| |
| return q.toString().trimEnd(); |
| } |
| |
| export function findSqlQueryPrefix(text: string): string | undefined { |
| try { |
| return extractQueryPrefix(text); |
| } catch (e) { |
| const startOffset = e.location?.start?.offset; |
| if (typeof startOffset !== 'number') return; |
| const prefix = text.slice(0, startOffset); |
| // Try to trim to where the error came from |
| try { |
| return extractQueryPrefix(prefix); |
| } catch { |
| // Try to trim out last word |
| try { |
| return extractQueryPrefix(prefix.replace(/\s*\w+$/, '')); |
| } catch { |
| return; |
| } |
| } |
| } |
| } |
| |
| export function cleanSqlQueryPrefix(text: string): string { |
| const matchReplace = /\sREPLACE$/i.exec(text); |
| if (matchReplace) { |
| // This query likely grabbed a "REPLACE" (which is not a reserved keyword) from the next query over, see if we can delete it |
| const textWithoutReplace = text.slice(0, -matchReplace[0].length).trimEnd(); |
| if (SqlQuery.maybeParse(textWithoutReplace)) { |
| return textWithoutReplace; |
| } |
| } |
| |
| return text; |
| } |
| |
| export interface QuerySlice { |
| index: number; |
| startOffset: number; |
| startRowColumn: RowColumn; |
| endOffset: number; |
| endRowColumn: RowColumn; |
| sql: string; |
| } |
| |
| export function findAllSqlQueriesInText(text: string): QuerySlice[] { |
| const found: QuerySlice[] = []; |
| |
| let remainingText = text; |
| let offset = 0; |
| let m: RegExpExecArray | null = null; |
| // Have an upper bound to how many queries we might extract from a long text |
| for (let i = 0; i < 1e3; i++) { |
| m = /SET|SELECT|WITH|INSERT|REPLACE|EXPLAIN/i.exec(remainingText); |
| if (!m) break; |
| |
| const sql = findSqlQueryPrefix(remainingText.slice(m.index)); |
| if (sql) { |
| const endIndex = m.index + sql.length; |
| found.push({ |
| index: found.length, |
| startOffset: offset + m.index, |
| startRowColumn: offsetToRowColumn(text, offset + m.index)!, |
| endOffset: offset + endIndex, |
| endRowColumn: offsetToRowColumn(text, offset + endIndex)!, |
| sql: cleanSqlQueryPrefix(sql), |
| }); |
| } |
| |
| let advanceBy = m.index; |
| const initialWord = m[0].toUpperCase(); |
| if (sql && initialWord === 'SET') { |
| const startOfSetStatements = remainingText.slice(m.index); |
| const [setPart, queryPart] = SqlSetStatement.partitionSetStatements( |
| startOfSetStatements, |
| true, |
| ); |
| advanceBy += setPart.length; |
| const nextWord = /SELECT|WITH|INSERT|REPLACE/i.exec(queryPart); |
| if (nextWord) { |
| advanceBy += nextWord[0].length; |
| } |
| } else { |
| advanceBy += initialWord.length; // Skip the initial word only |
| } |
| |
| remainingText = remainingText.slice(advanceBy); |
| offset += advanceBy; |
| } |
| |
| return found; |
| } |