Re: ERROR: wrong varnullingrels (b 5 7) (expected (b)) for Var 3/3 - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: ERROR: wrong varnullingrels (b 5 7) (expected (b)) for Var 3/3
Date
Msg-id ZGcPwqj1nd79XetH@paquier.xyz
Whole thread Raw
In response to ERROR: wrong varnullingrels (b 5 7) (expected (b)) for Var 3/3  (tender wang <tndrwang@gmail.com>)
Responses Re: ERROR: wrong varnullingrels (b 5 7) (expected (b)) for Var 3/3
List pgsql-hackers
On Fri, May 19, 2023 at 09:42:35AM +0800, tender wang wrote:
> postgres=# \d
> Did not find any relations.
> postgres=# CREATE TEMP TABLE t0(c0 inet, c1 money, c2 TEXT);
> CREATE TABLE
> postgres=# CREATE TEMP TABLE IF NOT EXISTS t1(LIKE t0);
> CREATE TABLE
> postgres=# CREATE TEMP TABLE t2(c0 boolean , c1 DECIMAL  NOT NULL UNIQUE);
> CREATE TABLE
> postgres=# CREATE TEMPORARY TABLE t3(LIKE t1);
> CREATE TABLE
> postgres=# CREATE VIEW v0(c0) AS (SELECT DISTINCT 0 FROM t3);
> NOTICE:  view "v0" will be a temporary view
> CREATE VIEW
> postgres=# SELECT SUM(count) FROM (SELECT
> (((t1.c2)LIKE(((((t0.c2)||(((103556691)-(v0.c0)))))||(v0.c0)))))::INT as
> count FROM t0, ONLY t1 RIGHT OUTER JOIN ONLY t2 ON t2.c0 RIGHT OUTER JOIN
> v0 ON ((t2.c1)=(0.08182310538090898))) as res;
> ERROR:  wrong varnullingrels (b 5 7) (expected (b)) for Var 3/3

Thanks for the test case, issue reproduced here on HEAD and not v15.
This causes an assertion failure here:
#4  0x000055a6f8faa776 in ExceptionalCondition (conditionName=0x55a6f915ac60 "bms_equal(rel->relids,
root->all_query_rels)",fileName=0x55a6f915ac3d "allpaths.c",  
    lineNumber=234) at assert.c:66
#5  0x000055a6f8c55b6d in make_one_rel (root=0x55a6fa814ea8, joinlist=0x55a6fa83f758) at allpaths.c:234
#6  0x000055a6f8c95c45 in query_planner (root=0x55a6fa814ea8, qp_callback=0x55a6f8c9c373 <standard_qp_callback>,
qp_extra=0x7ffc98138570)at planmain.c:278 
#7  0x000055a6f8c9860a in grouping_planner (root=0x55a6fa814ea8, tuple_fraction=0) at planner.c:1493

I am adding an open item.
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: createuser --memeber and PG 16
Next
From: Michael Paquier
Date:
Subject: Re: Conflict between regression tests namespace & transactions due to recent changes