Hello Gunther, That's an unfortunate regression. Thank you for the detailed bug report! I filed this issue here: https://sourceforge.net/apps/trac/jooq/ticket/886
> As a workaround, I am using a patched version of the AbstractSelect > class where the asTable(String alias) method does not delegate to the > no-arg variant but creates the SelectQueryAsTable directly. This > avoids the call to hashCode() and the exception. I think it also makes > more sense this way because the current code generates an alias > unnecessarily. Unfortunately, this alias is necessary in some database dialects, if I remember correctly. But it makes sense to review that code again. > But I was wondering: is there a newly imposed restriction on nested > selects? Am I doing it wrong? Your sample looks correct to me. I have very similar test cases (unfortunately, without dialect-specific functions) in https://github.com/lukaseder/jOOQ/blob/master/jOOQ-test/src/org/jooq/test/jOOQAbstractTest.java - testLimitNested() - testSubSelect() I can tell you more when I have reproduced the problem. Cheers Lukas
