ARRAY

On this page Carat arrow pointing down
Warning:
CockroachDB v22.1 is no longer supported. For more details, see the Release Support Policy.

The ARRAY data type stores one-dimensional, 1-indexed, homogeneous arrays of any non-array data type.

The ARRAY data type is useful for ensuring compatibility with ORMs and other tools. However, if such compatibility is not a concern, it's more flexible to design your schema with normalized tables.

CockroachDB supports indexing array columns with GIN indexes. This permits accelerating containment queries (@> and <@) on array columns by adding an index to them.

Note:

CockroachDB does not support nested arrays.

Syntax

A value of data type ARRAY can be expressed in the following ways:

  • Appending square brackets ([]) to any non-array data type.
  • Adding the term ARRAY to any non-array data type.

Size

The size of an ARRAY value is variable, but it's recommended to keep values under 1 MB to ensure performance. Above that threshold, write amplification and other considerations may cause significant performance degradation.

Functions

For the list of supported ARRAY functions, see Functions and Operators.

Examples

Creating an array column by appending square brackets

icon/buttons/copy
> CREATE TABLE a (b STRING[]);
icon/buttons/copy
> INSERT INTO a VALUES (ARRAY['sky', 'road', 'car']);
icon/buttons/copy
> SELECT * FROM a;
        b
------------------
  {sky,road,car}
(1 row)

Creating an array column by adding the term ARRAY

icon/buttons/copy
> CREATE TABLE c (d INT ARRAY);
icon/buttons/copy
> INSERT INTO c VALUES (ARRAY[10,20,30]);
icon/buttons/copy
> SELECT * FROM c;
      d
--------------
  {10,20,30}
(1 row)

Accessing an array element using array index

Note:

Arrays in CockroachDB are 1-indexed.

icon/buttons/copy
> SELECT * FROM c;
      d
--------------
  {10,20,30}
(1 row)
icon/buttons/copy
> SELECT d[2] FROM c;
  d
------
  20
(1 row)

Accessing an array column using containment queries

You can use the operators <@ ("is contained by") and @> ("contains") to run containment queries on ARRAY columns.

icon/buttons/copy
> SELECT * FROM c WHERE d <@ ARRAY[10,20,30,40,50];
      d
--------------
  {10,20,30}
(1 row)
icon/buttons/copy
> SELECT * FROM c WHERE d @> ARRAY[10,20];
      d
--------------
  {10,20,30}
(1 row)

Using the overlaps operator

You can use the && (overlaps) operator to select array columns by checking if another array overlaps the column array. Arrays overlap if they have any elements in common.

  1. Create the table:

    icon/buttons/copy
    CREATE TABLE a (b STRING[]);
    
  2. Insert two new arrays:

    icon/buttons/copy
    INSERT INTO a VALUES (ARRAY['runway', 'houses', 'city', 'clouds']);
    INSERT INTO a VALUES (ARRAY['runway', 'houses', 'city']);
    INSERT INTO a VALUES (ARRAY['sun','moon']);
    
  3. Use the && operator in a WHERE clause to a query:

    icon/buttons/copy
    SELECT * FROM a WHERE b && ARRAY['clouds','moon'];
    
                      b
    -------------------------------
      {runway,houses,city,clouds}
      {sun,moon}
    (2 rows)
    
    Time: 30ms total (execution 2ms / network 28ms)
    

Appending an element to an array

Using the array_append function

icon/buttons/copy
> SELECT * FROM c;
      d
--------------
  {10,20,30}
(1 row)
icon/buttons/copy
> UPDATE c SET d = array_append(d, 40) WHERE d[3] = 30;
icon/buttons/copy
> SELECT * FROM c;
        d
-----------------
  {10,20,30,40}
(1 row)

Using the append (||) operator

icon/buttons/copy
> SELECT * FROM c;
        d
-----------------
  {10,20,30,40}
(1 row)
icon/buttons/copy
> UPDATE c SET d = d || 50 WHERE d[4] = 40;
icon/buttons/copy
> SELECT * FROM c;
         d
--------------------
  {10,20,30,40,50}
(1 row)

Ordering by an array

icon/buttons/copy
> CREATE TABLE t (a INT ARRAY, b STRING);
icon/buttons/copy
> INSERT INTO t VALUES (ARRAY[3,4],'threefour'),(ARRAY[1,2],'onetwo');
icon/buttons/copy
> SELECT * FROM t;
    a   |     b
--------+------------
  {3,4} | threefour
  {1,2} | onetwo
(2 rows)
icon/buttons/copy
> SELECT * FROM t ORDER BY a;
    a   |     b
--------+------------
  {1,2} | onetwo
  {3,4} | threefour
(2 rows)

Supported casting and conversion

Casting between ARRAY values is supported when the data types of the arrays support casting. For example, it is possible to cast from a BOOL array to an INT array but not from a BOOL array to a TIMESTAMP array:

icon/buttons/copy
> SELECT ARRAY[true,false,true]::INT[];
   array
-----------
  {1,0,1}
(1 row)
icon/buttons/copy
> SELECT ARRAY[true,false,true]::TIMESTAMP[];
pq: invalid cast: bool[] -> TIMESTAMP[]

You can cast an array to a STRING value, for compatibility with PostgreSQL:

icon/buttons/copy
> SELECT ARRAY[1,NULL,3]::string;
    array
--------------
  {1,NULL,3}
(1 row)
icon/buttons/copy
> SELECT ARRAY[(1,'a b'),(2,'c"d')]::string;
               array
------------------------------------
  {"(1,\"a b\")","(2,\"c\"\"d\")"}
(1 row)

Implicit casting

CockroachDB supports implicit casting from string literals to arrays of all data types except the following:

For example, if you create a table with a column of type INT[]:

icon/buttons/copy
> CREATE TABLE x (a UUID DEFAULT gen_random_uuid() PRIMARY KEY, b INT[]);

And then insert a string containing a comma-delimited set of integers contained in brackets:

icon/buttons/copy
> INSERT INTO x(b) VALUES ('{1,2,3}'), (ARRAY[4,5,6]);

CockroachDB implicitly casts the string literal as an INT[]:

icon/buttons/copy
> SELECT * FROM x;
                   a                   |    b
---------------------------------------+----------
  2ec0ed91-8a82-4f2e-888e-ae86ece4fc60 | {4,5,6}
  a521d6e9-3a2a-490d-968c-1365cace038a | {1,2,3}
(2 rows)

See also


Yes No
On this page

Yes No