Home Page for the TeradataForum
 

Archives of the TeradataForum

Message Posted: Mon, 19 May 2003 @ 20:08:39 GMT


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


Subj:   Re: Size of count(*) field
 
From:   Anomy Anom

<-- Anonymously Posted: Monday, May 19, 2003 15:47 -->

This issue opens a can of worms. We discovered the count(*) limitation about 8 years ago, I think, and the details have escaped me now. I believe that utilities also run into a limitation doing counts. But even though the counters overflow, the SQL and the utilities still continue to work. They don't stop when they hit the limit. We've learned to live with it. The count(*) limit is an integer 2,147,483,647, but some processes seem to have a "count limit" that is twice that now (we're at 4.1.3). I'm sorry that I'm fuzzy on the details of what has changed and what's still supposed to be fixed someday.

By the way, last year NCR created a new warning message (via DR62648) to make this situation "clearer" to the user; to explain that an SQL process did run okay:

Insert into DB1.tbl1
select bla bla bla.
*** Insert completed. 4294967295 rows added.

*** Warning 6813 Numeric overflow in internal counters. The request is successful.

 SELECT COUNT(*) (DEC(18))
   FROM DB1.tbl1 ;
            Count(*)
--------------------
         5132041445.

In this case, the count of rows inserted into an empty table overflowed internally and stopped at 4.29 billion, but the insert/select actually did insert 5.13 billion rows successfully, which can be seen in the select count(*).



     
  <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: 27 Dec 2016