Hi All,
Im bulk loading a ton of data into MSSQL SERVER 2005 Standard Edition. I used to do this process in version 2000. It seems there is some more overhead in 2005. Is there a way to drop logging to almost null to speed up insert?
This is my current sql statment to load data.
EXEC sp_dboption 'my_stuff', 'select into/bulkcopy', 'true'
SET ANSI_WARNINGS OFF
BULK INSERT mystuff.dbo.[v1]
FROM 'c:\myfile.txt'
WITH
(
FIRSTROW = 1,
FORMATFILE = 'c:\scripts\v1.fmt',
MAXERRORS=2000,
ROWS_PER_BATCH=100000
)
Thanks,
Mike
There should not be any significant difference in what gets logged for this between SQL 2000 and SQL 2005. How are you measuring the discrepency?|||
I used to load this data on a windows 2000 server running sql server 2000. And I could load all of the data into 7-8 hours. Im running 22 scripts that load from 8 files a piece and specfic start and end points in the fmt files. The fmt also jump around through the data. The data files are 25 GB a piece. But now with the New version on a more powerful box the time esimates are going to be days. I guess I need to start looking for other variables if its not the logging.
I should also mention that the data all resides on a NetApp san and we are connected over 4GB fibre.
Thanks,
Mike
No comments:
Post a Comment