Hi,

I have a filter definition with a clause that looks something like this (for CloudScape):

"CURRENT_TIMESTAMP < TABLENAME_EFFECTIVITY_END"

This filter works as expected: the CURRENT_TIMESTAMP is inserted literally in the WHERE clause since it's a SQL function and the TABLENAME_EFFECTIVITY_END is correctly interpreted as a column name.

However, when I change the function to "CURRENT TIMESTAMP" so DB2 7.2 will understand it (with a space instead of an underscore), Hibernate 3.05 mistakes the CURRENT for a column name and generates a WHERE clause that looks something like

where prefix0_.CURRENT TIMESTAMP < prefix0_.TABLENAME_EFFECTIVITY_END

Thus, DB2 can't parse the resulting SQL.

I need a way to tell Hibernate to leave the "CURRENT TIMESTAMP" text alone since it's a SQL function, and to not interpret the first part as a (non-existent) column.
Is there currently such a way (docs don't mention it) and if not, could this be added as a feature? Maybe by placing the text between back ticks in the definition, like in other parts of the mapping file?

Joris

It might be worth submitting this to JIRA, since we could special-case it in the SQL parser.

A workaround would be to create a UDF called current_timestamp() and use that instead.

OK, I filed an issue. I registered it as a bug, though it could be considered an improvement:
http://opensource.atlassian.com/projects/hibernate/browse/HHH-736

I'll have a go with the user-defined function for now.

Tnx,

Joris