home clear 64x64
en blue 200x116 de orange 200x116 info letter User
suche 36x36
Latest versionsfixlist
11.1.0.7 FixList
10.5.0.9 FixList
10.1.0.6 FixList
9.8.0.5 FixList
9.7.0.11 FixList
9.5.0.10 FixList
9.1.0.12 FixList
Have problems? - contact us.
Register for free anmeldung-x26
Contact form kontakt-x26

DB2 - Problem description

Problem IC71982 Status: Closed

NESTED-LOOP JOIN WITH EARLYOUT FOR GROUPBY CLAUSES, YIELDS INCORRECT
RESULTS WHEN JOIN COLUMNS ARE OF DIFFERENT DATA TYPES

product:
DB2 FOR LUW / DB2FORLUW / 950 - DB2
Problem description:
When a query that meets all the following conditions is 
executed, fewer results may be returned: 
- Several tables are joined using columns with different data 
types 
- Nested-loop join is selected as the join method 
- The output from the nested-loop join method is included into 
the Group by clause and the Early Out flag is marked as GROUPBY 
 
The following scenarios indicate the problem: 
 
1. Joining two tables with different data types. To demonstrate 
this problem, create two tables for the test.  Create the tables 
such that the columns being joined (t1.i1 and t2.i1) have 
different data types. 
create table t1 (i1 bigint not null, c2 char(250)) 
create index i1 on t1 (i1) 
create table t2 (i1 smallint not null, i2 bigint) 
create index i2 on t2 (i1,i2) 
2. Run the query with different data types (bigint and smallint) 
in predicate and the result goes directly into the Group by 
clause. 
select t2.i1, max(t2.i2) from t1, t2 where t1.i1=t2.i1 group by 
t2.i1 
3. When the Nested-loop join is selected as the join method. 
 
Access Plan: 
----------- 
        Total Cost:             22.8806 
        Query Degree:           1 
              Rows 
             RETURN 
             (   1) 
              Cost 
               I/O 
               | 
               99 
             GRPBY 
             (   2) 
             22.8709 
                3 
               | 
               99 
             ^NLJOIN 
             (   3) 
             22.8655 
                3 
         /-----+------\ 
       99             12.1111 
     IXSCAN           IXSCAN 
     (   4)           (   5) 
    0.0455361         7.57676 
        0                1 
       |                | 
       99              1199 
 INDEX: DB2INST1   INDEX: DB2INST1 
       I1               I2 
       Q2               Q1 
        3) NLJOIN: (Nested Loop Join) 
                Arguments: 
                --------- 
                EARLYOUT: (Early Out flag) 
                        GROUPBY 
                Predicates: 
                ---------- 
                3) Predicate used in Join 
                        Predicate Text: 
                        -------------- 
                        (Q2."I1" = Q1."I1") 
 
In addition, the following message may be seen in the 
db2diag.log: 
 
2011-02-09-18.07.54.671541+060 I891882A2516       LEVEL: Severe 
PID     : 4567040              TID  : 23716       PROC : db2sysc 
1 
INSTANCE: instdb               NODE : 001         DB   : SAMPLE 
APPHDL  : 1-48010              APPID: APP1 
AUTHID  : AUTH1 
EDUID   : 23716                EDUNAME: db2agent (SAMPLE) 1 
FUNCTION: DB2 UDB, data services, sqlrxcon, probe:219 
DATA #1 : String, 117 bytes 
ASSERTION FAILED!!! 
ASSERTION EXPRESSION: convertNumRc == 0 
SOURCE FILE NAME: sqlrxcon.C 
SOURCE FILE LINE NUMBER: 208 
DATA #2 : Codepath, 8 bytes 
0 
DATA #3 : String, 72 bytes 
CONTACT DB2 SERVICE, assert a numeric conversion has 
unexpectedly failed 
DATA #4 : String, 13 bytes 
convertNumRc: 
DATA #5 : Hex integer, 4 bytes 
0x80160008 
... 
 
with the following stack: 
 
CALLSTCK: 
  [0] 0x0900000005FDA0BC pdLog + 0xEC 
  [1] 0x0900000005B3E91C sqlrxcon2__FP8SQLRXNLSP10sqlz_valueT2 + 
0x210 
  [2] 0x0900000005B845CC 
@147@keypartnd__FP7SQLI_CBP10sqlz_valuePPcP17COMP_FUNC_ELEMENTPl 
+ 0x24 
  [3] 0x09000000068CB03C 
sqliComparePrefixCompPg__FP7SQLI_CBP14SQLI_PAGE_DESCP8SQLD_KEYP1 
2SQLI_KEYDATAP9SQLI_SLOTPlP14COMP_FUNC_VECTPc 
+ 0xAF8 
  [4] 0x09000000068D43EC 
sqliComparePrefixCompPg__FP7SQLI_CBP14SQLI_PAGE_DESCP8SQLD_KEYP1 
2SQLI_KEYDATAP9SQLI_SLOTPlP14COMP_FUNC_VECTPc@glue634 
+ 0x74 
  [5] 0x0900000005F2784C 
sqliCompareGeneric__FP7SQLI_CBP14SQLI_PAGE_DESCP8SQLD_KEYP12SQLI 
_KEYDATAP9SQLI_SLOTPlP14COMP_FUNC_VECTPc 
+ 0x4 
  [6] 0x09000000068E6898 
sqliSearchFromCache__FP7SQLI_CBP11SQLI_SFGLOBPUl + 0x46C 
  [7] 0x090000000608A0E0 
.sqlirdk.fdpr.clone.2009__FP8sqeAgentP10SQLI_IXPCRP9SQLD_IXCBUlP 
8SQLD_KEYP12SQLI_KEYDATAP9sqli_scanPi 
+ 0x50 
  [8] 0x0900000005F67564 
.sqldIndexFetch.fdpr.clone.290__FP8SQLD_CCBP8SQLZ_RID + 0x1A4 
  [9] 0x0900000005F67D9C 
sqldRowFetch__FP8sqeAgentP8SQLD_CCBUlT3PP10SQLD_VALUEP8SQLZ_RIDT 
3P12SQLD_ID_LISTP9SQLP_LSN8 
+ 0x55C 
 
Under some conditions, the instance may trap.
Problem Summary:
**************************************************************** 
* USERS AFFECTED:                                              * 
* All users                                                    * 
**************************************************************** 
* PROBLEM DESCRIPTION:                                         * 
* When a query that meets all the following conditions is      * 
* executed, fewer results may be returned:                     * 
* - Several tables are joined using columns with different     * 
* data types                                                   * 
* - Nested-loop join is selected as the join method            * 
* - The output from the nested-loop join method is included    * 
* into the Group by clause and the Early Out flag is marked as * 
* GROUPBY                                                      * 
*                                                              * 
* The following scenarios indicate the problem:                * 
*                                                              * 
* 1. Joining two tables with different data types. To          * 
* demonstrate this problem, create two tables for the test.    * 
* Create the tables such that the columns being joined (t1.i1  * 
* and t2.i1) have different data types.                        * 
* create table t1 (i1 bigint not null, c2 char(250))           * 
* create index i1 on t1 (i1)                                   * 
* create table t2 (i1 smallint not null, i2 bigint)            * 
* create index i2 on t2 (i1,i2)                                * 
* 2. Run the query with different data types (bigint and       * 
* smallint) in predicate and the result goes directly into the * 
* Group by clause.                                             * 
* select t2.i1, max(t2.i2) from t1, t2 where t1.i1=t2.i1 group * 
* by t2.i1                                                     * 
* 3. When the Nested-loop join is selected as the join method. * 
*                                                              * 
* Access Plan:                                                 * 
* -----------                                                  * 
*         Total Cost:             22.8806                      * 
*         Query Degree:           1                            * 
*               Rows                                           * 
*              RETURN                                          * 
*              (   1)                                          * 
*               Cost                                           * 
*                I/O                                           * 
*                |                                             * 
*                99                                            * 
*              GRPBY                                           * 
*              (   2)                                          * 
*              22.8709                                         * 
*                 3                                            * 
*                |                                             * 
*                99                                            * 
*              ^NLJOIN                                         * 
*              (   3)                                          * 
*              22.8655                                         * 
*                 3                                            * 
*          /-----+------\                                      * 
*        99             12.1111                                * 
*      IXSCAN           IXSCAN                                 * 
*      (   4)           (   5)                                 * 
*     0.0455361         7.57676                                * 
*         0                1                                   * 
*        |                |                                    * 
*        99              1199                                  * 
*  INDEX: DB2INST1   INDEX: DB2INST1                           * 
*        I1               I2                                   * 
*        Q2               Q1                                   * 
*         3) NLJOIN: (Nested Loop Join)                        * 
*                 Arguments:                                   * 
*                 ---------                                    * 
*                 EARLYOUT: (Early Out flag)                   * 
*                         GROUPBY                              * 
*                 Predicates:                                  * 
*                 ----------                                   * 
*                 3) Predicate used in Join                    * 
*                         Predicate Text:                      * 
*                         --------------                       * 
*                         (Q2."I1" = Q1."I1")                  * 
**************************************************************** 
* RECOMMENDATION:                                              * 
* Upgrade to DB2 UDB Version 9.5 fix pack 7.                   * 
**************************************************************** 
Users affected: All users. 
Problem Description: Using the nested-loop join method, with 
earlyout indicator for groupby clauses, yields incorrect 
results when join columns are of different data types. 
. 
The problem occurs when the nested-loop join method, along with 
EARLYOUT indicator for groupby clauses, is used with a join 
clause that joins two columns with different data types. 
.
Local Fix:
Execute the following and then recycle the instance: 
db2set DB2_REDUCED_OPTIMIZATION=NO_NLJN_EO_FOR_GB
available fix packs:
DB2 Version 9.5 Fix Pack 8 for Linux, UNIX, and Windows
DB2 Version 9.5 Fix Pack 9 for Linux, UNIX, and Windows
DB2 Version 9.5 Fix Pack 10 for Linux, UNIX, and Windows

Solution
This problem was first fixed in DB2 UDB Version 9.5 fix pack 7.
Workaround
not known / see Local fix
BUG-Tracking
forerunner  : APAR is sysrouted TO one or more of the following: IC73581 IC73582 
follow-up : 
Timestamps
Date  - problem reported    :
Date  - problem closed      :
Date  - last modified       :
18.10.2010
05.01.2011
18.02.2011
Problem solved at the following versions (IBM BugInfos)
9.5.FP7
Problem solved according to the fixlist(s) of the following version(s)
9.1.0.7 FixList
9.5.0.7 FixList