Christopher Kings-Lynne <[EMAIL PROTECTED]> writes:
> Why doesn't this work:
> test=# select oid, relname, indisclustered from pg_index join pg_class
> on indexrelid=oid where indexrelid > 17205;
> ERROR: column "oid" does not exist
The JOIN is a bit like a view --- it doesn't have any system columns.
For example consider
select alias.* from (a join b on ...) as alias;
This should certainly not show any system columns --- both because SQL92
says so, and because there'd be no way to choose which input table's
system columns to use.
In scenarios where only one input table has an OID column, we could
maybe kluge things to allow it to be accessed without qualification,
but I don't really see the value of adding a wart for that ...
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
