StringTrimBoth¶

Overview¶

StringTrimBoth is a runtime replaceable expression that removes specified characters from both the beginning and end of a string. It serves as a wrapper around the StringTrim expression and provides the SQL function btrim for bilateral string trimming operations.

Syntax¶

btrim(string_expr [, trim_chars])

// DataFrame API usage through SQL expression
col("string_column").expr("btrim(string_column, ' ')")

Arguments¶

Argument	Type	Description
srcStr	Expression	The source string expression to be trimmed
trimStr	Option[Expression]	Optional expression specifying characters to trim. If not provided, whitespace characters are trimmed

Return Type¶

Returns a String type containing the trimmed result.

Supported Data Types¶

String types for the source string argument
String types for the optional trim characters argument

Algorithm¶

Inherits the trimming logic from the underlying StringTrim expression through runtime replacement
When trimStr is provided, removes all occurrences of those characters from both ends of the source string
When trimStr is omitted, defaults to trimming whitespace characters from both ends
Operates as a RuntimeReplaceable expression, delegating actual execution to StringTrim
Supports both single-argument (whitespace trimming) and two-argument (custom character trimming) variants

Partitioning Behavior¶

Preserves partitioning as it operates on individual string values without requiring data redistribution
Does not require shuffle operations since trimming is applied row-by-row

Edge Cases¶

Returns null when the source string is null
Returns null when the trim characters argument is null
Returns empty string when the source string contains only the characters specified for trimming
When trim string is empty, returns the original source string unchanged

Code Generation¶

As a RuntimeReplaceable expression, StringTrimBoth delegates code generation to its replacement expression (StringTrim), which supports Tungsten code generation for optimal performance.

Examples¶

-- Trim whitespace from both ends
SELECT btrim('  hello world  ');
-- Result: 'hello world'

-- Trim specific characters from both ends  
SELECT btrim('xxxhello worldxxx', 'x');
-- Result: 'hello world'

-- Example from source code documentation
SELECT btrim(encode('SSparkSQLS', 'utf-8'), encode('SL', 'utf-8'));
-- Result: 'parkSQ'

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

// Trim whitespace
df.select(expr("btrim(column_name)"))

// Trim specific characters
df.select(expr("btrim(column_name, 'abc')"))