Hyukjin Kwon created SPARK-61690:
------------------------------------
Summary: Make the simple string of the ML vector and matrix types
'vector' and 'matrix', as PySpark's VectorUDT and MatrixUDT
Key: SPARK-61690
URL: https://issues.apache.org/jira/browse/SPARK-61690
Project: Spark
Issue Type: Bug
Components: Connect
Affects Versions: connect-rust-4.3.0
Reporter: Hyukjin Kwon
The simple string of the ML vector and matrix types is "udt" in the Rust
client: `DataType::simple_string` of a user-defined type gives "udt". PySpark's
VectorUDT and MatrixUDT override `simpleString` to give "vector" and "matrix",
which is what schema strings of ML DataFrames show (e.g.
`struct<features:vector>`). Every other user-defined type stays "udt", and the
JSON and `typeName` forms are unchanged ("udt"), as in PySpark.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]