From: | Walter Cai <wzcai92(at)gmail(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-hackers(at)lists(dot)postgresql(dot)org |
Subject: | Re: |
Date: | 2018-01-09 20:58:11 |
Message-ID: | CAGCAsriVmBnnEOTJLCn6qYu=s2baC6AioqXF=f-7ikoR=YKq7w@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Sorry about not including a title at first, I completely forgot!
And thanks for the help, Tom. With respect to the Var: I'm really just
hoping to access the column name
Best, Walter
On Tue, Jan 9, 2018 at 11:14 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Walter Cai <wzcai92(at)gmail(dot)com> writes:
> > I (a graduate student) am currently trying to modify a postgres instance
> > slightly to incorporate my modified cardinality estimates. In order to
> run
> > these experiments I'm hoping to access the raw values for selections
> > predicates from within the calc_joinrel_size_estimate method (in
> costsize.c).
> > For example, if the restriction
>
> > WHERE name_attr = "example_str"
>
> > appears in the query, I'd like to get the "name_attr" and "example_str"
> > values. If possible, I would very much appreciate the help.
>
> It's pretty unclear what you mean by "raw values". What you're actually
> going to be dealing with in that part of the code is a List of
> RestrictInfo nodes, one for each relevant WHERE clause. The one
> representing this particular clause would contain an OpExpr node
> representing the "=" operator, and the two inputs of the operator
> would be a Var node representing the name_attr column and a Const
> representing the 'example_str' literal. You could pull the value
> of the literal out of the Const node easily enough, but I don't
> understand what you're looking for with respect to the Var.
>
> regards, tom lane
>
From | Date | Subject | |
---|---|---|---|
Next Message | Robert Haas | 2018-01-09 21:00:39 | Re: [HACKERS] UPDATE of partition key |
Previous Message | Tom Lane | 2018-01-09 20:54:10 | Re: improve type conversion of SPI_processed in Python |