count(9)Quote:
Originally posted by marist89
...I heard count(1) was faster... ;)
Printable View
count(9)Quote:
Originally posted by marist89
...I heard count(1) was faster... ;)
Ya sure...Even I know that not to be the case...:)Quote:
Originally posted by marist89
...I heard count(1) was faster... ;)
So, we don't have anything more efficient than a SELECT count(*)??
count(0)
Ya .. some DBAz working/worked on 7.x still think that :D :DQuote:
Originally posted by marist89
...I heard count(1) was faster... ;)
could you display the execution plan to check if an index is used for the count or if it does a fts
Count(NULL) shoud go even faster ;)Quote:
Originally posted by marist89
...I heard count(1) was faster... ;)
Once you go above 400 million, does an extra million really matter?
Just tell them it is a "big mother f*&%ing table"
MH
;)
You could always export the table and see how many rows it exported...
how about creating a nice function-based, compressed index ...
That ought to be reasonably small and quick to modify.Code:create index my_index on my_table ('X') pctfree 0 compress;
what about a procedure that runs when the system is quiet and populates a count history table (table_name,sysdate,count(*)) they can then just look at that. Its also pretty useful for spotting growth trends and what not.