dinesh_deadman Warnings : 1 New User
Joined: 10 Aug 2007 Posts: 32 Location: Chicago
|
|
|
|
Hello:
I have just created or rather changed an Index Controlled Partitioning to Table Controlled Partitioning by dropping partitioning index. And I expected the limit keys to be as below.
Code: |
---------+---------+---------+---------+---------+---------+---------+---------
TSNAME DBNAME LIMITKEY
---------+---------+---------+---------+---------+---------+---------+---------
TABLESPC DATABASE 'FM '
TABLESPC DATABASE 'ZZZ' |
For, my CREATE INDEX statement was like this before I decided to drop, in order to change PARTIONING TO Table Controlled...
Code: |
CREATE TYPE 2 UNIQUE INDEX TST.TESTINDX
ON TST.TEST_TABLE
(COL_FIRST ASC
,COL_SECOND ASC
.
.
.
,COL_LAST ASC )
CLUSTER
(PARTITION 1 ENDING AT ('FM ')
USING STOGROUP SYSDEFLT
PRIQTY 720
SECQTY 720
ERASE NO
FREEPAGE 0
PCTFREE 15
,PARTITION 2 ENDING AT ('ZZZ')
USING STOGROUP SYSDEFLT
PRIQTY 720
SECQTY 720
ERASE NO
FREEPAGE 0
.
.
. |
After dropping this INDEX I received a warning 20272 CODE (this is expected) saying the tablespace has been converted to Table-Controlled Partitioning instead of Index-Controlled Partitioning.
Now, I try to look at the limit keys in the catalog table here is what I see.
Code: |
---------+---------+---------+---------+---------+---------+---------+---------
TSNAME DBNAME LIMITKEY
---------+---------+---------+---------+---------+---------+---------+---------
TABLESPC DATABASE 'FM '
TABLESPC DATABASE MAXVALUE,MAXVALUE,M |
Does anyone have an idea why is it MAXVALUE,MAXVALUE... instead of 'ZZZ'???
DB2 V8 NFM - z/OS
Please let me know if I owe you any more information. |
|
sushanth bobby
Senior Member
Joined: 29 Jul 2008 Posts: 1020 Location: India
|
|
|
|
Dinesh,
From the manuals
Quote: |
After the conversion to table-controlled partitioning, DB2 changes the existing high-limit key value for non-large table spaces to the highest value for the key. Beginning in Version 8, DB2 enforces the high-limit key value. By default, DB2 does not put the last partition of the table space into a REORG-pending (REORP) state |
Sushanth |
|