lincoln-lil commented on PR #23478:
URL: https://github.com/apache/flink/pull/23478#issuecomment-1766456638
> 1. When constructing "non-comparable types," there is no problem with
`testSqlApi("NULL = f30", "NULL")`, but `testSqlApi("NULL = f30", "NULL")`
throws an exception:
> `Caused by: org.apache.calcite.sql.validate.SqlValidatorException:
Cannot apply '=' to arguments of type '<NULL> =
<RECORDTYPE:PEEK_NO_EXPAND(VARCHAR(2147483647) F0, VARCHAR(2147483647) F1)>'.
Supported form(s): '<COMPARABLE_TYPE> = <COMPARABLE_TYPE>'`
> 2. When constructing "non-comparable types," I haven't figured out how to
cover the branch:
> `} else if (isReference(right.resultType)) {` because the preceding
> `if (isReference(left.resultType)) {` always takes precedence.
> 3. I have written an SQL test for MULTISET in the MiscITCase. I'm not sure
if this is appropriate.
@fengjiajie IIUC, we should add an explicit type `cast` to `null` so that
the field type can be deterministic, e.g., 'cast(null as bigint)' to construct
a empty bigint column. For the `multiset` type, I'm fine with the current
`MiscITCase` or the `CalcItCase`.
Btw, during the community review process, we generally recommend using the
append commits, which makes it easier for the reviewer to track changes during
the process and squash them when merging at the end.
cc @LadyForest If you have time to help check it again that would be great :)
--
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: issues-unsubscr...@flink.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
