Обсуждение: UNION ALL - Var attno

sri harsha <sriharsha9992@gmail.com> writes:
>    Assume the following query ,
> (SELECT a * 1 , b FROM TABLE_1) UNION ALL (SELECT a *1 , b FROM TABLE_2);

> In this query , attribute number of the VARs are 141 and 2 respectively !!

I doubt it.

Maybe you're looking at something that's not a Var, possibly an OpExpr,
but assuming it's a Var?  The fact that 141 is the pg_proc OID of int4mul
lends considerable weight to this suspicion ...
        regards, tom lane

On Fri, Apr 29, 2016 at 2:42 PM, sri harsha <sriharsha9992@gmail.com> wrote:
>
> Its not an OpExpr . It is a VAR , got it from "reltargetlist" in base
> relation ( RelOptInfo) . Can you shed some light on where the conversion
> from 141 to "original" attribute number takes place ??

As Tom said, you must be looking at an OPEXPR's opfuncid value.
Because that's what I see as being 141.

Thanks,
Amit

sri harsha <sriharsha9992@gmail.com> writes:
> Its not an OpExpr . It is a VAR , got it from "reltargetlist" in base
> relation ( RelOptInfo) .

Read the comment:
*        reltargetlist - List of Var and PlaceHolderVar nodes for the values*                        we need to output
fromthis relation.*                        List is in no particular order, but all rels of an*
appendrelset must use corresponding orders.*                        NOTE: in an appendrel child relation, may contain*
                     arbitrary expressions pulled up from a subquery!
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

In general, it's seldom a good idea to assume that a Node is of a specific
type without having confirmed that with an IsA() check.
        regards, tom lane