Greg Hansen created SPARK-58917:
-----------------------------------
Summary: Respect inferSchema for variant ingestion in CSV parser
Key: SPARK-58917
URL: https://issues.apache.org/jira/browse/SPARK-58917
Project: Spark
Issue Type: Bug
Components: SQL
Affects Versions: 4.4.0
Reporter: Greg Hansen
The CSV to Variant parser (used by the `singleVariantColumn` option and
explicit `VariantType` columns) always infers scalar types (long, decimal,
date, timestamp, boolean) regardless of the `inferSchema` option, so a value
like "0001" is stored as the integer 1 rather than the string "0001".
A new CSV read option `variantRespectInferSchema` makes the parser honor
`inferSchema`: when it is set to true and `inferSchema` is false, scalar CSV
values are preserved as strings inside the Variant instead of being inferred.
The option defaults to false, preserving the existing always-infer behavior so
existing results are unchanged; users opt in per read. A reader option allows
users to set this on-read via `.option(...)`.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
