Re: parallelize queries containing subplans

On Mon, Jan 16, 2017 at 9:13 PM, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
>
> After commit-ab1f0c8, this patch needs a rebase. Attached find
> rebased version of the patch.
>
> Thanks to Kuntal for informing me offlist that this patch needs rebase.

In this patch, I have observed some changes while creating subplan for
CTE. So I have reviewed this from that perspective and also tried to
perform some test.

Here is my finding/ comments.

@@ -1213,6 +1216,7 @@ SS_process_ctes(PlannerInfo *root)
&splan->firstColCollation);
splan->useHashTable = false;
splan->unknownEqFalse = false;
+ splan->parallel_safe = best_path->parallel_safe;

I noticed that if path for CTE is parallel safe then we are marking
CTE subplan as parallel safe, In particular, I don't have any problem
with that, but can you add some test case which can cover this path, I
mean to say where CTE subplan are pushed.

------------
I have tried to test the subplan with CTE below is my test.
create table t1(a int , b varchar);
create table t (n int, b varchar);

Query:
explain verbose select * from t where t.n not in (WITH RECURSIVE t(n) AS (
VALUES (1)
UNION ALL
SELECT a+1 FROM t1 WHERE a < 100
)
SELECT sum(n) FROM t);

During debugging I found that subplan created for below part of the
query is parallel_unsafe, Is it a problem or there is some explanation
of why it's not parallel_safe,

(WITH RECURSIVE t(n) AS (
VALUES (1)
UNION ALL
SELECT a+1 FROM t1 WHERE a < 100
)
SELECT sum(n) FROM t);
----------

--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com