Hyukjin Kwon created SPARK-61012:
------------------------------------
Summary: Sample by the fractions of sampleBy, and take a Column,
in the Python client as PySpark does
Key: SPARK-61012
URL: https://issues.apache.org/jira/browse/SPARK-61012
Project: Spark
Issue Type: Bug
Components: Connect
Affects Versions: connect-rust-4.3.0
Reporter: Hyukjin Kwon
In PySpark Connect, {{df.sampleBy(col, fractions, seed)}} sends each stratum of
{{fractions}} with its fraction in the {{StatSampleBy}} relation, of a column
name or a Column {{col}}; it raises {{NOT_EXPECTED_TYPE}} for a {{col}} that is
neither or {{fractions}} that is not a dict, and
{{DISALLOWED_TYPE_FOR_CONTAINER}} for a stratum that is not a float, an int or
a string.
The Rust-backed Python client sends the strata without their fractions, so
every fraction is 0 and the sample is empty: {{df.sampleBy("k", {"a": 1.0, "b":
0.0}, seed=1).count()}} is 0, where PySpark 4.3.0 returns the 2 rows of
{{"a"}}. It also takes only a column name. Sample by the fractions as PySpark
does.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
