pyspark.sql.functions.collect_set#

pyspark.sql.functions.collect_set(col)[source]#

Aggregate function: Collects the values from a column into a set, eliminating duplicates, and returns this set of objects.

New in version 1.6.0.

Changed in version 3.4.0: Supports Spark Connect.

Parameters

colColumn or column name: The target column on which the function is computed.

Returns

Column: A new Column object representing a set of collected values, duplicates excluded.

See also

pyspark.sql.functions.array_agg()
pyspark.sql.functions.collect_list()

Notes

This function is non-deterministic as the order of collected results depends on the order of the rows, which may be non-deterministic after any shuffle operations.

Examples

Example 1: Collect values from a DataFrame and sort the result in ascending order

>>> from pyspark.sql import functions as sf
>>> df = spark.createDataFrame([(1,), (2,), (2,)], ('value',))
>>> df.select(sf.sort_array(sf.collect_set('value')).alias('sorted_set')).show()
+----------+
|sorted_set|
+----------+
|    [1, 2]|
+----------+

Example 2: Collect values from a DataFrame and sort the result in descending order

>>> from pyspark.sql import functions as sf
>>> df = spark.createDataFrame([(2,), (5,), (5,)], ('age',))
>>> df.select(sf.sort_array(sf.collect_set('age'), asc=False).alias('sorted_set')).show()
+----------+
|sorted_set|
+----------+
|    [5, 2]|
+----------+

Example 3: Collect values from a DataFrame with multiple columns and sort the result

>>> from pyspark.sql import functions as sf
>>> df = spark.createDataFrame([(1, "John"), (2, "John"), (3, "Ana")], ("id", "name"))
>>> df = df.groupBy("name").agg(sf.sort_array(sf.collect_set('id')).alias('sorted_set'))
>>> df.orderBy(sf.desc("name")).show()
+----+----------+
|name|sorted_set|
+----+----------+
|John|    [1, 2]|
| Ana|       [3]|
+----+----------+