DanielLeens commented on PR #12802:
URL: https://github.com/apache/seatunnel/pull/12802#issuecomment-6851084150

   Thanks @SEZ9. The requested `.xls` fast-path regression is already on 
`6e554c48`: `ExcelReadStrategyTest.testXlsReadKeepsNumericPrecision` (829-861) 
is parameterized over `POI` and `EasyExcel`, creates an actual HSSFWorkbook 
`.xls`, caches a `1/3` formula result, reads it through the selected engine, 
and asserts the stored DOUBLE, DECIMAL and formula values. 
`ExcelCellUtils.java:118-122` uses `originalNumberValue` only for a numeric 
cell with a non-null original value. This identifies the regression covering 
the claim; I have not executed it in this review round.
   
   For the remaining ledger: the numeric-cell semantics are documented in 
`docs/en/connectors/source/LocalFile.md:367`; the sink's mixed numeric/text 
DECIMAL behavior and the existing SQL `CAST(... AS DOUBLE/STRING)` alternatives 
are documented in `docs/en/connectors/sink/LocalFile.md:200`, so a new 
all-numeric sink option is deliberately not added. Declared scale is applied 
without rounding only when it fits (`ExcelCellUtils.java:149-168`). The 
large-exponent tests and their non-BigDecimal conversion comment remain in 
`ExcelGeneratorTest`; please use these current-head pointers to finish your 
independent reconciliation. The prior source conclusion is unchanged. 
Current-head Build evidence is still required; an older green head is not 
acceptance for this SHA, and no empty commit is needed just to request a manual 
Build.


-- 
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]

Reply via email to