Home Page for the TeradataForum
 

Archives of the TeradataForum

Message Posted: Thu, 18 Jan 2001 @ 16:12:12 GMT


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


Subj:   Re: OR in a WHERE clause
 
From:   Rudel Simard

Thank you for your suggestion Claybourne, we try to figure out an other way to create single report than do 2 separate reports. It will be less expensive for the 'developer report maintenance'.

I understand the way Teradata does for the utilisation of the index and the statistics on the tables are good. I use EXPLAIN often to understand the way Teradata works with the NUPI, statistics... I give you, Frank, the 2 explains to help your understanding of my problem.

I really appreciate your help, thank for your support

     EXPLAIN
     SELECT T1."COL1_NUSI" (NAMED "c1" ) ,
      T2."COL2_NUSI" (NAMED "c2" ) FROM
     ( DONNE.TABLE2 T2 LEFT OUTER JOIN DONNE.TABLE1 T1 ON
     T2."COL3_NUPI" = T1."COL3_NUPI")

     WHERE  (1 = 1 AND 25113 = T2."COL2_NUSI" OR
      1 = 2 AND '123' = T1."COL1_NUSI"  ) and
      T2.COL4_NUSI BETWEEN 990101 AND 991231;

      *** Help information returned. 29 rows.
      *** Total elapsed time was 1 second.

Explanation
--------------------------------------------------
 
  1)First, we lock a distinct DONNE."pseudo table" for read on a RowHash to prevent global deadlock for DONNE.T1.  
  2)Next, we lock a distinct DONNE."pseudo table" for read on a RowHash to prevent global deadlock for DONNE.T2.  
  3)We lock DONNE.T1 for read, and we lock DONNE.T2 for read.  
  4)We do an all-AMPs JOIN step from DONNE.T1 by way of a RowHash match scan with no residual conditions, which is joined to DONNE.T2 with a condition of ("((DONNE.T2.COL2_NUSI = 25113) OR (1 = 2 )) AND ((DONNE.T2.COL4_NUSI <= DATE '1999-12-31') AND ((DONNE.T2.COL4_NUSI >= DATE '1999-01-01') AND ((1 = 2) OR (1 = 1 ))))"). DONNE.T1 and DONNE.T2 are right outer joined using a merge join, with a join condition of ("DONNE.T2.COL3_NUPI = DONNE.T1.COL3_NUPI"). The input tables DONNE.T1 and DONNE.T2 will not be cached in memory. The result goes into Spool 2, which is built locally on the AMPs. The result spool file will not be cached in memory. The size of Spool 2 is estimated with low confidence to be 63,308,472 rows. The estimated time for this step is 1 hour and 9 minutes.  
  5)We do an all-AMPs RETRIEVE step from Spool 2 (Last Use) by way of an all-rows scan with a condition of ("((1 = 1) OR (Spool_2.COL1_NUSI = '123')) AND ((Spool_2.COL2_NUSI = 25113) OR (Spool_2.COL1_NUSI = '123'))") into Spool 1, which is built locally on the AMPs. The result spool file will not be cached in memory. The size of Spool 1 is estimated with low confidence to be 63,308,472 rows. The estimated time for this step is 6 minutes and 32 seconds.  
  6)Finally, we send out an END TRANSACTION step to all AMPs involved in processing the request.  
  ->The contents of Spool 1 are sent back to the user as the result of statement 1. The total estimated time is 1 hour and 15 minutes.  


     EXPLAIN
     SELECT T1."COL1_NUSI" (NAMED "c1" ) ,
      T2."COL2_NUSI" (NAMED "c2" ) FROM
     ( TABLE2 T2 LEFT OUTER JOIN TABLE1 T1 ON
     T2."COL3_NUPI" = T1."COL3_NUPI")

     WHERE  (1 = 1 AND 25113 = T2."COL2_NUSI") and
      T2.COL4_NUSI BETWEEN 990101 AND 991231;

      *** Help information returned. 20 rows.
      *** Total elapsed time was 1 second.

Explanation
--------------------------------------------------
 
  1)First, we lock a distinct DONNE."pseudo table" for read on a RowHash to prevent global deadlock for DONNE.T1.  
  2)Next, we lock a distinct DONNE."pseudo table" for read on a RowHash to prevent global deadlock for DONNE.T2.  
  3)We lock DONNE.T1 for read, and we lock DONNE.T2 for read.  
  4)We do an all-AMPs JOIN step from DONNE.T2 by way of index # 20 "DONNE.T2.COL2_NUSI = 25113" with a residual condition of ("(DONNE.T2.COL4_NUSI <= DATE '1999-12-31') AND ((DONNE.T2.COL4_NUSI >= DATE '1999-01-01') AND (1 = 1 ))"), which is joined to DONNE.T1. DONNE.T2 and DONNE.T1 are left outer joined using a merge join, with a join condition of ("DONNE.T2.COL3_NUPI = DONNE.T1.COL3_NUPI"). The input tables DONNE.T2 and DONNE.T1 will not be cached in memory. The result goes into Spool 1, which is built locally on the AMPs. The size of Spool 1 is estimated with low confidence to be 24,844 rows. The estimated time for this step is 2 minutes and 7 seconds.  
  5)Finally, we send out an END TRANSACTION step to all AMPs involved in processing the request.  
  ->The contents of Spool 1 are sent back to the user as the result of statement 1. The total estimated time is 2 minutes and 7 seconds.  


Rudel Simard
R�gie de l'assurance-maladie du Qu�bec
DBA de l'environnement informationnel



     
  <Prev Next>   <<First <Prev Next> Last>>  
 
 
 
 
 
 
 
 
  
  Top Home Privacy Feedback  
 
 
Copyright for the TeradataForum (TDATA-L), Manta BlueSky    
Copyright 2016 - All Rights Reserved    
Last Modified: 15 Jun 2023