Philip Warner <pjw@rhyme.com.au> writes:
> At 03:06 AM 20/10/2004, Tom Lane wrote:
>> I think the tricky part of that would be inserting the tablespace clause
>> in the right place; for CREATE INDEX this seems to require nontrivial
>> parsing. (Both the index column definitions and the WHERE clause could
>> be arbitrarily complicated expressions.) If we can get around that part
>> then this wouldn't be too hard.
> I may be missing something here; I was assuming that pg_dump would dump
> would build the CREATE INDEX/TABLE/etc commands with the %%tablespace%%
> already embedded. pg_restore would not need to do any parsing. Or is there
> something I don't understand?
Maybe there's something I don't understand. How are you expecting
pg_restore to control whether it outputs the command with a TABLESPACE
clause embedded or not, if pg_dump has already built the command string
that way? I thought you were envisioning that pg_restore would insert,
or not insert, a TABLESPACE clause into a command that didn't initially
have one.
regards, tom lane