Re: [Postgres-xc-bugs] Query are shipped to the data nodes more than needed.

[Ashutosh B. <ash...@en...>]

Hi Julian,
I looked at the patch and tried to apply the patch, but there is an error
[ashutosh@ubuntu coderoot]git apply
/mnt/hgfs/tmp/20140110_more_datanode_involved.patch
/mnt/hgfs/tmp/20140110_more_datanode_involved.patch:10: trailing whitespace.
        if (and_clause((Node *) qual_expr))
/mnt/hgfs/tmp/20140110_more_datanode_involved.patch:11: trailing whitespace.
        {
/mnt/hgfs/tmp/20140110_more_datanode_involved.patch:12: trailing whitespace.
            List *sub_lquals = ((BoolExpr *) qual_expr)->args;
/mnt/hgfs/tmp/20140110_more_datanode_involved.patch:13: trailing whitespace.
            distcol_expr = pgxc_find_distcol_expr(varno, attrNum,
sub_lquals);
/mnt/hgfs/tmp/20140110_more_datanode_involved.patch:14: trailing whitespace.
            if (distcol_expr != NULL)
fatal: corrupt patch at line 20

Did you use git diff for creating the patch?

BTW, I looked at the patch. The patch seems to be applied to function
pgxc_find_distcol_expr() (functions name usually appears along with the
diff, if you are using git diff).

In this function there is code

828     /* Convert the qualification into List if it's not already so */
829     if (!IsA(quals, List))
830         lquals = make_ands_implicit((Expr *)quals);
831     else
832         lquals = (List *)quals;

which takes care of the quals which are not in the List form. This part of
the code should have taken care of the case you mentioned. It seems that
for some reason, this part is not converting the ANDed quals into List. Can
you please check why? May be you want to step through make_ands_implicit()
(using gdb or some debugger) to see what's happening there.




On Fri, Jan 10, 2014 at 3:38 PM, ZhangJulian <jul...@ou...>wrote:

>
> Your name                  : Julian ZL Zhang
> Your email address        : jul...@ou...
>
>
> System Configuration:
> ---------------------
>   Architecture (example: Intel Pentium)                : Intel Pentium
>
>   Operating System (example: Linux 2.4.18)        : 2.6.32-358.el6.x86_64
>
>   Postgres-XC version (example: Postgres-XC 1.1devel): Github master
>   Compiler used (example: gcc 3.3.5)                : gcc (GCC) 4.4.7
> 20120313 (Red Hat 4.4.7-3)
>
>
> Please enter a FULL description of your problem:
> ------------------------------------------------
> I had opened the bug at http://sourceforge.net/p/postgres-xc/bugs/467/ just
> descripte it more detailedly here.
> Different predicate order results in different plan, the query should be
> shipped only to one datanode, but in the first example, two data nodes are
> involved.
>
> Please describe a way to repeat the problem. Please try to provide a
> concise reproducible example, if at all possible:
> ----------------------------------------------------------------------
>  postgres=# create table t1 (c1 int, c2 int, c3 int) distribute by
> hash(c1);
> CREATE TABLE
> postgres=# explain verbose select * from t1 where c1=1 and c2=1 and c3=1;
> QUERY PLAN
>
> --------------------------------------------------------------------------------
> Data Node Scan on "REMOTE_FQS_QUERY" (cost=0.00..0.00 rows=0 width=0)
> Output: t1.c1, t1.c2, t1.c3
> Node/s: datanode1, datanode2
> Remote query: SELECT c1, c2, c3 FROM benchmarksql.t1 WHERE (((c1 = 1) AND
> (c2 = 1)) AND (c3 = 1))
> (4 rows)
> postgres=# explain verbose select * from t1 where c2=1 and c3=1 and c1=1;
> QUERY PLAN
>
> --------------------------------------------------------------------------------
> Data Node Scan on "REMOTE_FQS_QUERY" (cost=0.00..0.00 rows=0 width=0)
> Output: t1.c1, t1.c2, t1.c3
> Node/s: datanode1
> Remote query: SELECT c1, c2, c3 FROM benchmarksql.t1 WHERE (((c2 = 1) AND
> (c3 = 1)) AND (c1 = 1))
> (4 rows)
>
>
> If you know how this problem might be fixed, list the solution below:
> ---------------------------------------------------------------------
> I made a fix as the append file, which can resolve this issue, but I am
> not sure if it is correct.
> When I run MAKE CHECK, 3 test cases failed. The results are appended too.
> But when I run MAKE CHECK without the fix, there are also some test cases
> failed.
> I do not know if the failed test cases are related to the fix.
>
> Thanks
> Julian
>
>
> ------------------------------------------------------------------------------
> CenturyLink Cloud: The Leader in Enterprise Cloud Services.
> Learn Why More Businesses Are Choosing CenturyLink Cloud For
> Critical Workloads, Development Environments & Everything In Between.
> Get a Quote or Start a Free Trial Today.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
> _______________________________________________
> Postgres-xc-bugs mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-bugs
>
>


-- 
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company