Tom Lane <[EMAIL PROTECTED]> writes:

> Greg is correct.  int8 is a pass-by-reference datatype 

Just to keep the conversation on track. the evidence from this particular post
seems to indicate that my theory was wrong and the overhead for count(*) is
_not_ a big time sink. It seems to be at most 10% and usually less. A simple
"select 1 from tab" takes nearly as long.

I'm still puzzled why the times on these are so different when the latter
returns fewer records and both are doing sequential scans:

 select 1 from tab

 select 1 from tab where a < 1000


---------------------------(end of broadcast)---------------------------
TIP 7: don't forget to increase your free space map settings

Reply via email to