ArraySize¶

Overview¶

The ArraySize expression returns the number of elements in an array. It is a runtime-replaceable expression that internally delegates to the Size expression with legacySizeOfNull set to false, providing consistent null handling behavior.

Syntax¶

array_size(array_expr)

// DataFrame API
col("array_column").expr("array_size(array_column)")

Arguments¶

Argument	Type	Description
child	Expression	The array expression whose size will be calculated

Return Type¶

Returns an IntegerType representing the number of elements in the array.

Supported Data Types¶

ArrayType - Arrays of any element type (string, numeric, complex types, etc.)

Algorithm¶

Evaluates the input array expression to get the array value
If the array is null, returns null (non-legacy behavior)
If the array is not null, counts the number of elements in the array
Returns the count as an integer value
Delegates the actual size calculation to the underlying Size expression

Partitioning Behavior¶

This expression preserves partitioning since it operates on individual rows without requiring data movement:

Does not require shuffle operations
Can be executed independently on each partition
Maintains existing data distribution

Edge Cases¶

Null arrays: Returns null when the input array is null (non-legacy behavior)
Empty arrays: Returns 0 for arrays with no elements
Nested arrays: Counts only the top-level elements, not nested array contents
Arrays with null elements: Null elements within the array are counted as regular elements

Code Generation¶

This expression supports code generation through its runtime replacement mechanism:

Uses the underlying Size expression's code generation capabilities
Benefits from Tungsten code generation optimizations
Falls back to interpreted mode if code generation fails

Examples¶

-- Basic array size calculation
SELECT array_size(array('b', 'd', 'c', 'a'));
-- Returns: 4

-- Empty array
SELECT array_size(array());
-- Returns: 0

-- Null array
SELECT array_size(NULL);
-- Returns: NULL

-- Array with null elements
SELECT array_size(array('a', NULL, 'c'));
-- Returns: 3

// DataFrame API usage
import org.apache.spark.sql.functions._

df.select(expr("array_size(array_column)"))

// With array creation
df.select(expr("array_size(array('a', 'b', 'c'))"))