Re: [sqlite] COUNT() on indexed tables /primary keywith 100'000records

2008-07-14 Thread Igor Tandetnik
<[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> I supposed a very fast query even with many records, because COUNT( )
> might do its work on the index only.

Well, it does, but the work is still proportional to the number of 
records counted. The index helps to quickly filter out records that 
_don't_ match the condition, but you still have to look at every record 
that does.

Igor Tandetnik 



___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users


Re: [sqlite] COUNT() on indexed tables / primary keywith 100'000records

2008-07-13 Thread [EMAIL PROTECTED]
Hello

Thank you for you quick responses.

If I disable the index with + it takes around 50% more execution time.

I supposed a very fast query even with many records, because COUNT( ) might do 
its work on the index only.

I this case, I have to find another solution to get the number of records 
quickly. The often discussed "trigger 
COUNT() solution" might not be applicable here.

Daniel

Ursprüngliche Nachricht
Von: [EMAIL PROTECTED]
Datum: 14.07.2008 00:51
An: 
Betreff: Re: [sqlite] COUNT() on indexed tables / primary   keywith 
100'000records

<[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> Lets say i have 120'000 records and worst case is that upto 90'000
> records match the conditions.

Well, if you count almost all records, why do you expect the query to 
run substantially faster than the one that in fact counts all records?

Igor Tandetnik



___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users



___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users


Re: [sqlite] COUNT() on indexed tables / primary keywith 100'000records

2008-07-13 Thread Igor Tandetnik
<[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> Lets say i have 120'000 records and worst case is that upto 90'000
> records match the conditions.

Well, if you count almost all records, why do you expect the query to 
run substantially faster than the one that in fact counts all records?

Igor Tandetnik



___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users