<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html
  PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />

<meta name="copyright" content="(C) Copyright 2025" />
<meta name="DC.rights.owner" content="(C) Copyright 2025" />
<meta name="DC.Type" content="concept" />
<meta name="DC.Title" content="ARRAY Complex Type (Impala 2.3 or higher only)" />
<meta name="DC.Relation" scheme="URI" content="../topics/impala_datatypes.html" />
<meta name="prodname" content="Impala" />
<meta name="prodname" content="Impala" />
<meta name="version" content="Impala 3.4.x" />
<meta name="version" content="Impala 3.4.x" />
<meta name="DC.Format" content="XHTML" />
<meta name="DC.Identifier" content="array" />
<link rel="stylesheet" type="text/css" href="../commonltr.css" />
<title>ARRAY Complex Type (Impala 2.3 or higher only)</title>
</head>
<body id="array">


  <h1 class="title topictitle1" id="ariaid-title1">ARRAY Complex Type (<span class="keyword">Impala 2.3</span> or higher only)</h1>


  <div class="body conbody">

    <p class="p">
      A complex data type that can represent an arbitrary number of ordered elements.
      The elements can be scalars or another complex type (<code class="ph codeph">ARRAY</code>,
      <code class="ph codeph">STRUCT</code>, or <code class="ph codeph">MAP</code>).
    </p>


    <p class="p">
        <strong class="ph b">Syntax:</strong>
      </p>


<pre class="pre codeblock"><code><var class="keyword varname">column_name</var> ARRAY &lt; <var class="keyword varname">type</var> &gt;

type ::= <var class="keyword varname">primitive_type</var> | <var class="keyword varname">complex_type</var>
</code></pre>

      <p class="p">
        <strong class="ph b">Usage notes:</strong>
      </p>


      <p class="p">
        Because complex types are often used in combination, for example an
        <code class="ph codeph">ARRAY</code> of <code class="ph codeph">STRUCT</code> elements, if you are unfamiliar with
        the Impala complex types, start with
        <a class="xref" href="../shared/../topics/impala_complex_types.html#complex_types">Complex Types (Impala 2.3 or higher only)</a> for background
        information and usage examples.
      </p>


      <p class="p">
        The elements of the array have no names. You refer to the value of the array item using the
        <code class="ph codeph">ITEM</code> pseudocolumn, or its position in the array with the <code class="ph codeph">POS</code>
        pseudocolumn. See <a class="xref" href="impala_complex_types.html#item">ITEM and POS Pseudocolumns</a> for information about
        these pseudocolumns.
      </p>


    <p class="p">
      Each row can have a different number of elements (including none) in the array for that row.
    </p>


      <p class="p">
        When an array contains items of scalar types, you can use aggregation functions on the array elements without using join notation. For
        example, you can find the <code class="ph codeph">COUNT()</code>, <code class="ph codeph">AVG()</code>, <code class="ph codeph">SUM()</code>, and so on of numeric array
        elements, or the <code class="ph codeph">MAX()</code> and <code class="ph codeph">MIN()</code> of any scalar array elements by referring to
        <code class="ph codeph"><var class="keyword varname">table_name</var>.<var class="keyword varname">array_column</var></code> in the <code class="ph codeph">FROM</code> clause of the query. When
        you need to cross-reference values from the array with scalar values from the same row, such as by including a <code class="ph codeph">GROUP
        BY</code> clause to produce a separate aggregated result for each row, then the join clause is required.
      </p>


      <p class="p">
        A common usage pattern with complex types is to have an array as the top-level type for the column:
        an array of structs, an array of maps, or an array of arrays.
        For example, you can model a denormalized table by creating a column that is an <code class="ph codeph">ARRAY</code>
        of <code class="ph codeph">STRUCT</code> elements; each item in the array represents a row from a table that would
        normally be used in a join query. This kind of data structure lets you essentially denormalize tables by
        associating multiple rows from one table with the matching row in another table.
      </p>


      <p class="p">
        You typically do not create more than one top-level <code class="ph codeph">ARRAY</code> column, because if there is
        some relationship between the elements of multiple arrays, it is convenient to model the data as
        an array of another complex type element (either <code class="ph codeph">STRUCT</code> or <code class="ph codeph">MAP</code>).
      </p>


      <p class="p">
        You can pass a multi-part qualified name to <code class="ph codeph">DESCRIBE</code> to specify an
        <code class="ph codeph">ARRAY</code>, <code class="ph codeph">STRUCT</code>, or <code class="ph codeph">MAP</code> column and
        visualize its structure as if it were a table. For example, if table <code class="ph codeph">T1</code>
        contains an <code class="ph codeph">ARRAY</code> column <code class="ph codeph">A1</code>, you could issue the
        statement <code class="ph codeph">DESCRIBE t1.a1</code>. If table <code class="ph codeph">T1</code> contained a
        <code class="ph codeph">STRUCT</code> column <code class="ph codeph">S1</code>, and a field <code class="ph codeph">F1</code>
        within the <code class="ph codeph">STRUCT</code> was a <code class="ph codeph">MAP</code>, you could issue the
        statement <code class="ph codeph">DESCRIBE t1.s1.f1</code>. An <code class="ph codeph">ARRAY</code> is shown as a
        two-column table, with <code class="ph codeph">ITEM</code> and <code class="ph codeph">POS</code> columns. A
        <code class="ph codeph">STRUCT</code> is shown as a table with each field representing a column in the
        table. A <code class="ph codeph">MAP</code> is shown as a two-column table, with <code class="ph codeph">KEY</code>
        and <code class="ph codeph">VALUE</code> columns.
      </p>


      <p class="p">
        <strong class="ph b">Added in:</strong> <span class="keyword">Impala 2.3.0</span>
      </p>


      <p class="p">
        <strong class="ph b">Restrictions:</strong>
      </p>


      <ul class="ul">
        <li class="li">
          <p class="p">
            Columns with this data type can only be used in tables or partitions with the
            Parquet or ORC file format.
          </p>

        </li>


        <li class="li">
          <p class="p">
            Columns with this data type cannot be used as partition key columns in a partitioned
            table.
          </p>

        </li>


        <li class="li">
          <p class="p">
            The <code class="ph codeph">COMPUTE STATS</code> statement does not produce any statistics for
            columns of this data type.
          </p>

        </li>


        <li class="li">
          <p class="p" id="array__d6e3305">
            The maximum length of the column definition for any complex type, including
            declarations for any nested types, is 4000 characters.
          </p>

        </li>


        <li class="li">
          <p class="p">
            See <a class="xref" href="../shared/../topics/impala_complex_types.html#complex_types_limits">Limitations and Restrictions for Complex Types</a> for a
            full list of limitations and associated guidelines about complex type columns.
          </p>

        </li>

      </ul>


      <p class="p">
        <strong class="ph b">Kudu considerations:</strong>
      </p>

      <p class="p"> Currently, the data types
        <code class="ph codeph">CHAR</code>, <code class="ph codeph">ARRAY</code>, <code class="ph codeph">MAP</code>, and
          <code class="ph codeph">STRUCT</code> cannot be used with Kudu tables.
      </p>


      <p class="p">
        <strong class="ph b">Examples:</strong>
      </p>


      <div class="note note"><span class="notetitle">Note:</span> 
        Many of the complex type examples refer to tables such as <code class="ph codeph">CUSTOMER</code> and
        <code class="ph codeph">REGION</code> adapted from the tables used in the TPC-H benchmark. See
        <a class="xref" href="../shared/../topics/impala_complex_types.html#complex_sample_schema">Sample Schema and Data for Experimenting with Impala Complex Types</a> for the table
        definitions.
      </div>


      <p class="p">
        The following example shows how to construct a table with various kinds of <code class="ph codeph">ARRAY</code> columns,
        both at the top level and nested within other complex types.
        Whenever the <code class="ph codeph">ARRAY</code> consists of a scalar value, such as in the <code class="ph codeph">PETS</code>
        column or the <code class="ph codeph">CHILDREN</code> field, you can see that future expansion is limited.
        For example, you could not easily evolve the schema to record the kind of pet or the child's birthday alongside the name.
        Therefore, it is more common to use an <code class="ph codeph">ARRAY</code> whose elements are of <code class="ph codeph">STRUCT</code> type,
        to associate multiple fields with each array element.
      </p>


      <div class="note note"><span class="notetitle">Note:</span> 
        Practice the <code class="ph codeph">CREATE TABLE</code> and query notation for complex type columns
        using empty tables, until you can visualize a complex data structure and construct corresponding SQL statements reliably.
      </div>


<pre class="pre codeblock"><code>CREATE TABLE array_demo
(
  id BIGINT,
  name STRING,
-- An ARRAY of scalar type as a top-level column.
  pets ARRAY &lt;STRING&gt;,

-- An ARRAY with elements of complex type (STRUCT).
  places_lived ARRAY &lt; STRUCT &lt;
    place: STRING,
    start_year: INT
  &gt;&gt;,

-- An ARRAY as a field (CHILDREN) within a STRUCT.
-- (The STRUCT is inside another ARRAY, because it is rare
-- for a STRUCT to be a top-level column.)
  marriages ARRAY &lt; STRUCT &lt;
    spouse: STRING,
    children: ARRAY &lt;STRING&gt;
  &gt;&gt;,

-- An ARRAY as the value part of a MAP.
-- The first MAP field (the key) would be a value such as
-- 'Parent' or 'Grandparent', and the corresponding array would
-- represent 2 parents, 4 grandparents, and so on.
  ancestors MAP &lt; STRING, ARRAY &lt;STRING&gt; &gt;
)
STORED AS PARQUET;

</code></pre>

    <p class="p">
      The following example shows how to examine the structure of a table containing one or more <code class="ph codeph">ARRAY</code> columns by using the
      <code class="ph codeph">DESCRIBE</code> statement. You can visualize each <code class="ph codeph">ARRAY</code> as its own two-column table, with columns
      <code class="ph codeph">ITEM</code> and <code class="ph codeph">POS</code>.
    </p>


<pre class="pre codeblock"><code>DESCRIBE array_demo;
+--------------+---------------------------+
| name         | type                      |
+--------------+---------------------------+
| id           | bigint                    |
| name         | string                    |
| pets         | array&lt;string&gt;             |
| marriages    | array&lt;struct&lt;             |
|              |   spouse:string,          |
|              |   children:array&lt;string&gt;  |
|              | &gt;&gt;                        |
| places_lived | array&lt;struct&lt;             |
|              |   place:string,           |
|              |   start_year:int          |
|              | &gt;&gt;                        |
| ancestors    | map&lt;string,array&lt;string&gt;&gt; |
+--------------+---------------------------+

DESCRIBE array_demo.pets;
+------+--------+
| name | type   |
+------+--------+
| item | string |
| pos  | bigint |
+------+--------+

DESCRIBE array_demo.marriages;
+------+--------------------------+
| name | type                     |
+------+--------------------------+
| item | struct&lt;                  |
|      |   spouse:string,         |
|      |   children:array&lt;string&gt; |
|      | &gt;                        |
| pos  | bigint                   |
+------+--------------------------+

DESCRIBE array_demo.places_lived;
+------+------------------+
| name | type             |
+------+------------------+
| item | struct&lt;          |
|      |   place:string,  |
|      |   start_year:int |
|      | &gt;                |
| pos  | bigint           |
+------+------------------+

DESCRIBE array_demo.ancestors;
+-------+---------------+
| name  | type          |
+-------+---------------+
| key   | string        |
| value | array&lt;string&gt; |
+-------+---------------+

</code></pre>

    <p class="p">
      The following example shows queries involving <code class="ph codeph">ARRAY</code> columns containing elements of scalar or complex types. You
      <span class="q">"unpack"</span> each <code class="ph codeph">ARRAY</code> column by referring to it in a join query, as if it were a separate table with
      <code class="ph codeph">ITEM</code> and <code class="ph codeph">POS</code> columns. If the array element is a scalar type, you refer to its value using the
      <code class="ph codeph">ITEM</code> pseudocolumn. If the array element is a <code class="ph codeph">STRUCT</code>, you refer to the <code class="ph codeph">STRUCT</code> fields
      using dot notation and the field names. If the array element is another <code class="ph codeph">ARRAY</code> or a <code class="ph codeph">MAP</code>, you use
      another level of join to unpack the nested collection elements.
    </p>


<pre class="pre codeblock"><code>-- Array of scalar values.
-- Each array element represents a single string, plus we know its position in the array.
SELECT id, name, pets.pos, pets.item FROM array_demo, array_demo.pets;

-- Array of structs.
-- Now each array element has named fields, possibly of different types.
-- You can consider an ARRAY of STRUCT to represent a table inside another table.
SELECT id, name, places_lived.pos, places_lived.item.place, places_lived.item.start_year
FROM array_demo, array_demo.places_lived;

-- The .ITEM name is optional for array elements that are structs.
-- The following query is equivalent to the previous one, with .ITEM
-- removed from the column references.
SELECT id, name, places_lived.pos, places_lived.place, places_lived.start_year
  FROM array_demo, array_demo.places_lived;

-- To filter specific items from the array, do comparisons against the .POS or .ITEM
-- pseudocolumns, or names of struct fields, in the WHERE clause.
SELECT id, name, pets.item FROM array_demo, array_demo.pets
  WHERE pets.pos in (0, 1, 3);

SELECT id, name, pets.item FROM array_demo, array_demo.pets
  WHERE pets.item LIKE 'Mr. %';

SELECT id, name, places_lived.pos, places_lived.place, places_lived.start_year
  FROM array_demo, array_demo.places_lived
WHERE places_lived.place like '%California%';

</code></pre>

    <p class="p">
        <strong class="ph b">Related information:</strong>
      </p>


    <p class="p">
      <a class="xref" href="impala_complex_types.html#complex_types">Complex Types (Impala 2.3 or higher only)</a>,

      <a class="xref" href="impala_struct.html#struct">STRUCT Complex Type (Impala 2.3 or higher only)</a>, <a class="xref" href="impala_map.html#map">MAP Complex Type (Impala 2.3 or higher only)</a>
    </p>


  </div>


<div class="related-links">
<div class="familylinks">
<div class="parentlink"><strong>Parent topic:</strong> <a class="link" href="../topics/impala_datatypes.html">Data Types</a></div>
</div>
</div></body>
</html>