Ignoring dot operator in Teradata BTEQ - Part 2

[MarkBell920]

This is following on from my previous question on trying to ignore BTEQ in Teradata SQL. Thank you for the help.

This works and I can catch the BTEQ keywords but this has caused another problem. Several of the BTEQ keywords are also valid Teradata SQL keywords so extending the Teradata dialect file breaks the SQL parsing. The most import BTEQ word here is SET.

So this would be valid Teradata SQL:

SET QUERYBAND = 'queryband details';

UPDATE table1
SET column1 = 12
WHERE column2 = 30;

With the dot operator on SET QUERYBAND being optional.

Any ideas on getting round this?

Mark

[georgesittas]

Hey, can you elaborate on what's the issue here? Is it that SET is no longer parsed into an exp.Set and it's an exp.Command?

[MarkBell920]

Yes that is the problem. Set is now treated like a command not part of the update.

Mark

[georgesittas]

Hm that's related to how _parse_update works, i.e. it expects a SET token instead of a COMMAND one.. Perhaps a viable alternative for your use case is to override the TokenType.SET entry in STATEMENT_PARSERS and have it call _parse_as_command instead of _parse_set, so you wouldn't have to map "SET" into TokenType.COMMAND in the Tokenizer, only ".SET"

[MarkBell920]

Thanks. Would that cause problems where the BTEQ command is not prefixed with a dot?

.SET and SET are both valid BTEQ commands.

Mark

[georgesittas]

Thanks. Would that cause problems where the BTEQ command is not prefixed with a dot?

No, but you'd lose fine-grained parsing of the SET statement because it'd be parsed into a Command expression.

[MarkBell920]

OK thanks. I will try that and see if it works.

Mark