Hyukjin Kwon created SPARK-60907:
------------------------------------
Summary: Send reader schemas in Spark Connect Rust as PySpark
does, and accept a StructType in schema()
Key: SPARK-60907
URL: https://issues.apache.org/jira/browse/SPARK-60907
Project: Spark
Issue Type: Bug
Components: Connect
Affects Versions: connect-rust-4.3.0
Reporter: Hyukjin Kwon
In PySpark Connect, {{DataFrameReader.schema}} and {{DataStreamReader.schema}}
take a StructType, which is sent as its JSON, or a string, which is sent as is;
anything else raises {{NOT_EXPECTED_TYPE}}. The {{schema}} arguments of
{{load()}}, {{csv()}}, {{json()}} and the other format methods go through the
same method.
The Rust-backed Python client differs:
* {{schema()}} accepts only a string, so {{spark.read.schema(struct_type)}} and
{{spark.readStream.schema(struct_type)}} raise {{TypeError}}.
* The {{schema}} arguments of the format methods are parsed on the client and
sent as a simple string. A StructType with field metadata, such as a comment,
then fails on the server with {{PARSE_SYNTAX_ERROR}}, and a DDL string that the
client parser does not accept, such as one with {{COMMENT}}, fails with
{{ValueError}}. Data types other than a StructType are accepted rather than
raising {{NOT_EXPECTED_TYPE}}.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
