Archives of the TeradataForum
Message Posted: Mon, 25 Oct 2004 @ 12:51:34 GMT
Am new to Teradata and need help on this fast. We are working on V2R4 40 amps. Currently have series of 12 transaction tables each devided by all the transaction for the month (containing about 80-100 million records). I have a view which combines all the 12 tables into one (union) and joins with the sys_calendar.calendar on the transaction date. On each of the 12 base tables I have a value ordered index on the call start date.
Now whenever I run a query for the month = '06' i.e June, instead of going to the particular table for '06' Teradata goes and scans all the tables for data related to 06 . The query takes a lot of time. I tried creating a constraint on the date for each of the monthly table. But that doesn't seem to have worked either. How can I force Teradata to look at only the table for which I have data instead of scanning all the tables.
What can I do differently? I was wondering if I should convert the value ordered index on the transaction date to to a simple NUSI as Teradata doesn't do range scan on the transaction tables now.
I was wondering is this way of doing horizontal rule based partitions a good idea. Has anybody attempted this?
Appreciate your help.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 23 Jun 2019|