srielau commented on PR #59188:
URL: https://github.com/apache/spark/pull/59188#issuecomment-6042486338
@dtenedor Thanks for the approval and for catching the comma-delimiter case.
Pinned the default-comma behavior you listed: `schema_of_csv(CAST('1' AS
CHAR(3)))` is `STRUCT<_c0: DOUBLE>`, `from_csv` on padded `'1,2'` / `'1,ab'`
keeps the last-field pad (INT becomes null, STRING keeps `"ab "`), and
`ignoreTrailingWhiteSpace` restores `{1, 2}`. Same checks are in
`CsvExpressionsSuite` for interpreted and generated eval. Also added a parquet
`CHAR(5)` column so the table-read path is covered, not only a folded CAST
literal.
Updated the description for the comma case, `ignoreTrailingWhiteSpace`, and
the `readSideCharPadding` / standard-semantics disagreement on CHAR columns.
On the smaller notes: past `maxColumns` a space-delimited wide CHAR fails
the same way as a STRING of that width, so a follow-up sentence is enough. The
XmlToStructs codegen/interpreted mix-up is not in the commit messages.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
