| Bug #9618 | Test 'group_by' returns more rows than expected when NDB is included | ||
|---|---|---|---|
| Submitted: | 4 Apr 2005 18:16 | Modified: | 6 Apr 2005 9:55 |
| Reporter: | Joerg Bruehe | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S2 (Serious) |
| Version: | 4.1.11 | OS: | x86: Linux, OpenBSD, QNX, Sol 8 |
| Assigned to: | Joerg Bruehe | CPU Architecture: | Any |
[4 Apr 2005 18:18]
Joerg Bruehe
Correction: 'qnx' has no working 'cluster', so here it is 'max' only.
[5 Apr 2005 15:01]
Sergey Petrunya
Reassigning to Joerg as I cannot repeat the problem
[6 Apr 2005 9:55]
Joerg Bruehe
Sorry - this was caused by a handling error. An old source/test package was used for the run, on a rebuild the problem was gone.
Description: Release build of 4.1.11, based on ChangeSet 1.2174 05/04/01 12:35:49 msvensson@neptunus.(none) +2 -0 Manual merge This difference occurs identical in the 'cluster' and 'max' builds on 'build' (Linux/i686), 'buildqnx', 'ds9' (OpenBSD 3.7 /i386), and 'intelxeon3' (Solaris 8 /i386) ('qnx' has no working 'max', so here it is 'cluster' only). ------------------------------------------------------- *** r/group_by.result Thu Mar 31 11:57:09 2005 --- r/group_by.reject Fri Apr 1 23:35:41 2005 *************** *** 699,704 **** --- 699,708 ---- 1 SIMPLE t2 ref PRIMARY,a PRIMARY 400 const,const 6 Using where select c from t2 where a = 2 and b = 'val-2' group by c; c + val-110 + val-38 + val-50 + val-62 val-74 val-98 drop table t1,t2; ------------------------------------------------------- How to repeat: Build + test on one of these.