[VL] Enable cast when input type equal to dest type.

[j7nhai]

What changes were proposed in this pull request?

(Please fill in changes proposed in this fix)

(Fixes: #ISSUE-ID)

How was this patch tested?

(Please explain how this patch was tested. E.g. unit tests, integration tests, manual tests)

(If this patch involves UI changes, please attach a screenshot; otherwise, remove this)

[github-actions]

Thanks for opening a pull request!

Could you open an issue for this pull request on Github Issues?

https://github.com/apache/incubator-gluten/issues

Then could you also rename commit message and pull request title in the following format?

[GLUTEN-${ISSUES_ID}][COMPONENT]feat/fix: ${detailed message}

See also:

• Other pull requests

[marin-ma]

This case shouldn't happen. See https://github.com/apache/incubator-gluten/blob/main/docs/velox-backend-support-progress.md#cast-functions-support-status

N/A: not applicable case, e.g., from type is as same as to type, where cast will not actually happen.

Could you provide a sample to demonstrate when this code path will be called?

[j7nhai]

This case shouldn't happen. See https://github.com/apache/incubator-gluten/blob/main/docs/velox-backend-support-progress.md#cast-functions-support-status

N/A: not applicable case, e.g., from type is as same as to type, where cast will not actually happen.

Could you provide a sample to demonstrate when this code path will be called?

sql("use test")
sql("drop table if exists t1")
sql("drop table if exists t2")

sql("create table t1 (a TIMESTAMP) USING ICEBERG")
sql("create table t2 (b TIMESTAMP) USING ICEBERG")

sql("INSERT OVERWRITE t2 VALUES (CAST('2025-01-21 00:00:00' AS TIMESTAMP))")
sql("INSERT OVERWRITE t1 SELECT * FROM t2")  // Not supported to map spark function name to substrait function name: cast(b#7 as timestamp), class name: AnsiCast.

Then I try to convert AnsiCast to Cast when origin type equal to dest type. And the native cast not support ts to ts.

[PHILO-HE]

@j7nhai, does this issue only happen for iceberg? I found using parquet has no issue.

Spark has an optimizer that removes the same type cast. Not sure why it is not applied in your case.
https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/expressions.scala#L1091