Archives of the TeradataForum

Message Posted: Tue, 18 Aug 2015 @ 19:31:32 GMT

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

Hello Forum - I am trying to tune the below query and I am out of options.

The below query runs over 2 hours. Could anyone give me a direction here?

It's doing an all rows scan due to the CASE logic and on top of that , I also need to remove duplicates also (that's why using a SET table).

More details below.

Underlying table count - 5.5 Billion records PI of the table - HCI,date The date column is also the PPI Stats are collected on columns - HCI,date and POP There are no 'Low' or 'no' confidence in EXPLAIN plan.

Below is POP column count grouped by A,B,C,D

          POP           Count(*)
     1    A           4,603,656,794.00
     2    B           10,544,361.00
     3    C           355,998,280.00
     4    D           548,883,167.00

Things I tried -

Created a secondary index on POP and it didn't help.

Tried splitting the query into two using UNION. That didn't help either

     CREATE VOLATILE SET TABLE MEDICAL AS
     (
     SELECT
         0 as ID,
         HCI,
         date,
         Type,
         optin,
         D1,
         D2,
         D3,
         D4,
         .
         .
         .
         .
         .
         CASE WHEN POP = 'A' THEN 1
             WHEN POP = 'B' THEN 2
             WHEN POP = 'C' THEN 3
             WHEN POP = 'D' THEN 5
         END AS Flag,
         1 As TFlag
     FROM viewMED
     WITH DATA PRIMARY INDEX (HCI,date,)
      ON COMMIT PRESERVE ROWS;

Thanks in advance for all the suggestions and inputs!

Roopalini

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