SELECT COUNT(*) FROM {tableName.Fullname} performance
badgerparade opened this issue · comments
If a table has an int primary key and has a large number of rows (e.g. > 100 million) then the SELECT COUNT(*) has a negative impact on performance because of the scan on the table. For me it's adding in 4-5 seconds per bulk insert.
Looking at the code it looks like the count is only used to see if there's any records in the table for setting the nextId. You could change from a "SELECT COUNT(*) FROM ...." to a "SELECT TOP 1 1 FROM ..." and then handle the null as a zero if the table doesn't have any records.
Or you could change the CommandText to be "SELECT CASE WHEN EXISTS (SELECT TOP 1 * FROM {tableName.Fullname}) THEN 1 ELSE 0 END"
That's a good idea. Thanks!
I just made the suggested change and started a rather long running db population program trying it out.
Cool, hopefully you'll see the speed improvement. If you ever want a big database for playing around with I love using the stack overflow public database dump. You can get the SQL Server version from here: https://www.brentozar.com/archive/2018/06/new-stack-overflow-public-database-available-2018-06/