|
From: Jason L. W. <jwh...@jw...> - 2007-07-25 13:21:49
|
> Dmitry Yemanov escreveu: > >> Now, for the first problem. I cannot use the sh.QTY in the > >> FOR UPDATE list. > >> I can't even use the fully qualified name of the table as a prefix. > >> I have to actually use the field name without any relation name or > >> alias at all. > > > > Right, this is accordingly to the SQL specification. What you need could > > be done in ten minutes, but we already have a few places in the parser > > marked with "IBO hack" (table aliases in INSERT and UPDATE, for example) > > :-) And the intention was to restore the standard syntax as soon as IBO > > is fixed. So, I really doubt that we need more such "hacks" now. > > > What you mean in "this is accordingly to the SQL specification"? > AFAIR, the spec allows table aliases in UPDATE and DELETE. > > And I see in UPDATE syntax: > /* CVC: This option should be deprecated! The only allowed syntax should > be > Update...set column = expr, without qualifier for the column. */ > > I think this comment is wrong. I agree that it is wrong. One reason is that without being able to specify a table alias that I won't be able to specifically distinguish columns in the positioned update statements. I probably should flesh one of these out as a sample to see exactly how it works. I've already worked all night and I probably better get some sleep at this point. Thanks again, Jason L Wharton |