Archives of the TeradataForum

Message Posted: Wed, 17 Aug 2005 @ 18:31:26 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Sharma,

Here is a way of deleting duplicate rows the way you want it...

1) Modify your table to have a PI that includes all columns

     CREATE MULTISET TABLE TEST.test1 ,NO FALLBACK ,
          NO BEFORE JOURNAL,
          NO AFTER JOURNAL,
          CHECKSUM = DEFAULT
          (
           a VARCHAR(20) CHARACTER SET LATIN NOT CASESPECIFIC,
           b VARCHAR(20) CHARACTER SET LATIN NOT CASESPECIFIC,
           c DECIMAL(17,0))
     PRIMARY INDEX ( a ,b ,c );

2) Enable DBSControl Internal Flag "ROWIDEnabled" to TRUE. This would allow you to use ROWID in your DML.

(I believe this is to be used very cautiously & in exceptional cases only. Don't have details to fill in here)

3) delete from test1 where substr(rowid,9,1) > '01'xb;

Regards,

Sanjaya

	<Prev	Next>		<<First	<Prev	Next>	Last>>