[SPARK-55636][CONNECT] Add detailed errors in case of deduplication of invalid columns

[pranavdev022]

What changes were proposed in this pull request?

This PR updates the error handling for invalid deduplicate column names in Spark Connect to use the standard UNRESOLVED_COLUMN_AMONG_FIELD_NAMES error class instead of throwing INTERNAL_ERROR, a generic error message.

Example	Classic	Connect (Before)	Connect (After)
	Cannot resolve column name "artist_id" among (id, song_name, artist_name).	[INTERNAL_ERROR] Invalid deduplicate column artist_id SQLSTATE: XX000	[UNRESOLVED_COLUMN_AMONG_FIELD_NAMES] Cannot resolve column name "artist_id" among (id, song_name, artist_name). SQLSTATE: 42703
	Cannot resolve column name "cont.f1" among (id, cont).	[INTERNAL_ERROR] Invalid deduplicate column cont.f1 SQLSTATE: XX000	[UNRESOLVED_COLUMN_AMONG_FIELD_NAMES] Cannot resolve column name "cont.f1" among (id, cont). SQLSTATE: 42703
	works	works	works
	Cannot resolve column name "song.names" among (id, song.name, artist_name).	[INTERNAL_ERROR] Invalid deduplicate column song.names SQLSTATE: XX000	[UNRESOLVED_COLUMN_AMONG_FIELD_NAMES] Cannot resolve column name "song.names" among (id, song.name, artist_name). SQLSTATE: 42703
	works	works	works
	Cannot resolve column name "cont.value" among (id, cont.val).	[INTERNAL_ERROR] Invalid deduplicate column cont.value SQLSTATE: XX000	[UNRESOLVED_COLUMN_AMONG_FIELD_NAMES] Cannot resolve column name "cont.value" among (id, cont.val). SQLSTATE: 42703
		same	same

Why are the changes needed?

The previous error message in Spark Connect was not consistent with classic Spark and lacked helpful context.
This change aligns Spark Connect error messages with classic Spark, providing users with:

1. The correct error class (UNRESOLVED_COLUMN_AMONG_FIELD_NAMES instead of INTERNAL_ERROR).
2. The correct SQLSTATE (42703 instead of XX000).
3. A list of available column names to help users fix the issue.

Does this PR introduce any user-facing change?

Yes. Error messages for invalid deduplicate column names in Spark Connect are now more detailed and consistent with classic Spark.

How was this patch tested?

Tested with a custom image with the proposed changes.

Was this patch authored or co-authored using generative AI tooling?

No

[khakhlyuk]

LGTM, thanks for the fix!

[khakhlyuk]

nitty nit: this line can be moved between 1448 and 1449 since it's not used in the outer scope

[pranavdev022]

yes moved, that would be better.

[holdenk]

Thanks for working on error messages/debugging/dev ex :)