xiedeyantu commented on PR #4547: URL: https://github.com/apache/calcite/pull/4547#issuecomment-3315980893
I'm getting more confused with the revisions in this PR. Should I unify the parameters of public APIs like `sortLimit`, `sort`, and l`imit` to Number, or should we keep the current approach? However, Julian mentioned in #4537 that it's better to use Number to represent FETCH and OFFSET. Based on this, I could pass Number when using APIs like `sortLimit`, `sort`, and `limit`. But should the definitions of these APIs themselves remain as long? (The original definitions used int, which is indeed insufficient, so I understand that modifying the public methods is acceptable.) My understanding is that `long` is practically sufficient in engineering practice, but as Julian mentioned, theoretically, we shouldn’t impose limitations. Would using BigInteger be appropriate? However, when extracting numbers from `RexLiteral`, BigDecimal is used instead. Comparing BigDecimal and BigInteger would require type conversion, which introduces additional complexity. If we use BigDecimal, we would be concerned about the input values being floating-point numbers. So, what should I do? Can anyone offer some suggestions? -- 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]
