Mark Buckle wrote: > > Ingres uses the same kind of syntax as given below, > Oracle uses the target column = (nested subquery) syntax. > Gives us plenty of scope for fun when porting Ingres to oracle and vice > versa.
And as I mentioned, PostgreSQL also uses the FROM clause on the UPDATE. > Which is standard ? Doesn't matter, you have to use the syntax your > database(s) support. In practical terms I agree with you that one should use the syntax supported by one's database, but I totally disagree that it "doesn't matter". I am working on SQL::Statement and I need to keep the ANSI supported syntax completely separate from the non-standard extenstions. The module will support both, but in different ways. And for exactly the reason you mentioned: "gives us plenty of scope for fun when porting Ingres to oracle ..." If there is ever going to be a tool that helps in that process (and I'm not claiming that SQL::Statement is yet such a tool, but it should aim high) then the distinction between what is an actual ANSI standard and what is implementation specific extension needs to be maintained. -- Jeff
