theta_difference

Returns the set difference of two binary representations of Datasketches ThetaSketch objects (elements in first sketch but not in second), using a Datasketches ANotB object.

Syntax

from pyspark.sql import functions as sf

sf.theta_difference(col1, col2)

Parameters

Parameter	Type	Description
col1	pyspark.sql.Column or str	The first Theta sketch.
col2	pyspark.sql.Column or str	The second Theta sketch.

Returns

pyspark.sql.Column: The binary representation of the difference ThetaSketch.

Examples

Example 1: Get difference of two Theta sketches

from pyspark.sql import functions as sf
df = spark.createDataFrame([(1,4),(2,4),(3,5),(4,5)], "struct<v1:int,v2:int>")
df = df.agg(
    sf.theta_sketch_agg("v1").alias("sketch1"),
    sf.theta_sketch_agg("v2").alias("sketch2")
)
df.select(sf.theta_sketch_estimate(sf.theta_difference(df.sketch1, "sketch2"))).show()

+---------------------------------------------------------+
|theta_sketch_estimate(theta_difference(sketch1, sketch2))|
+---------------------------------------------------------+
|                                                        3|
+---------------------------------------------------------+