anybody experience with grouping datatables with a lot of rows (50k)? My IPhone 3G 开发者_高级运维takes 11seconds!!! for executing a group by statement.
2011-06-11 09:20:50.719 crmclient[1040:307] Begin Select count(*) as count, substr(upper(Pers_LastName),1,1) as identifier From contact_sync2 WHERE 1=1 GROUP BY substr(upper(Pers_LastName),1,1) order by Pers_LastName COLLATE NOCASE
2011-06-11 09:20:50.736 crmclient[1040:307] End
2011-06-11 09:21:05.047 crmclient[1040:307] RSEnd
2011-06-11 09:21:05.534 crmclient[1040:307] Begin Select Pers_PersonId, Pers_LastName || ', ' || Pers_FirstName AS Pers_FullName From contact_sync2 WHERE 1=1 order by Pers_LastName COLLATE NOCASE LIMIT 0,140
2011-06-11 09:21:05.546 crmclient[1040:307] End
2011-06-11 09:21:13.426 crmclient[1040:307] RSEnd
The first one gives me the structure of the whole table to show on iphone. The second gives me the first 140 rows. RSEnd is printed when the sqlite.step is done.
Thank you very much. I think 50.000 rows is not that much.
Maybe is better in your case to keep in a separate column the first letter. In this way you could eliminate the 2 functions (substr and upper) from your query as these takes time.
The select will be:
Select
count(*) as count, FirstLetter
From
contact_sync2
GROUP BY
FirstLetter
ORDER BY
Pers_LastName COLLATE NOCASE
Also put an index on FirstLetter.
If you can't create indexes how fast would it be without using ORDER BY, and then sorting using Objective-C?
精彩评论