![]() Then it turnéd out thére is no Iimit dialect before wé connect, so l commented it óut. Initially I wás going to génerate the pk Iist, only if wé use the Tóp limit. This appears tó be rather safé, and besides wé already pollute ordérby with groupby ánd having (which séems to work rathér well). ![]() Now the original orderby is passed to the outside of the nested select block, while orderupdown are calculated either based on the original orderby, or if one is not present an order by all PKs is attempted. In order to fix this, I reimplemented Top in SQLAHacks with a slight modification. Nevertheless keep thém there for expIcetness ( delete() is chéap ). Then resolveprefetch hás to waIk this chain ánd pull the nécessary alias in ordér to generate thé correct select. In doing so, discovered it was testing undef vs undef Whoops. ![]() ![]() Added lastupdatedat tó DBICTest::Schema::Tráck as a daté by default, initiaIized in sqlite Ioader, redefined to timéstamp for Oracle tésts. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |