thswlsqls opened a new pull request, #10738:
URL: https://github.com/apache/paimon/pull/10738
### Purpose
fix #10737
- `SparkProcedureUtils` normalizes each `partitions` value with the Spark
SQL parser and keeps `literalValue.toString()`, so an unquoted `hh=08` on a
STRING partition becomes `8`; procedures using this path then target partition
`8` (or nothing) and still return `true`.
- For STRING/VARCHAR/CHAR partition columns, only quoted string literals are
unquoted now; other values are kept as written. Other column types and
`parseStaticPartition` are unchanged.
- Follow-up to #8390, which reused the #7313 overwrite-spec normalizer for
procedure partitions. Flink already passes the raw value.
- Out of scope: `compact_chain_table` already fails with
`ClassCastException` for unquoted STRING values on master (separate follow-up).
### Tests
- Added `CompactProcedureTestBase` "compact partitions keep leading zeros of
string values" (STRING `hh=08`, STRING `hh="08"`, INT `hh=08`). On master it
fails: `{08}` keeps 2 files while `{8}` is compacted.
- `CompactProcedureTest` (Spark 3.5, `-Pspark3`): 53/53 passed. Spark 4
(`-Pspark4`) was not run locally.
--
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]
