Copy large amount of records via stored procedure?
Hello,
I have a view which returns close to 5 million rows of data. I need to copy this data into a table of the same structure.
Is it possible to create a stored procedure which copies the data over in chunks (so it doesn't timeout)?
If so can you give an example or a link to a tutorial?
Thanks,
-Torrwin
I have a view which returns close to 5 million rows of data. I need to copy this data into a table of the same structure.
Is it possible to create a stored procedure which copies the data over in chunks (so it doesn't timeout)?
If so can you give an example or a link to a tutorial?
Thanks,
-Torrwin
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
A "one off" is an event you only plan to do once.
http://www.vbrad.com/article.aspx?id=10
Explains the benefits of bcp over using stored procedures (speed) and has some vb class/exe that maybe you could include in your project?
http://www.vbrad.com/article.aspx?id=10
Explains the benefits of bcp over using stored procedures (speed) and has some vb class/exe that maybe you could include in your project?
ASKER
No, this will happen several times a month.
Also, BCP appears to need the source to be a file, not a view.
Also, BCP appears to need the source to be a file, not a view.
select * into newTable from MyView
will transfer all rows into newTable, but 5 million rows is relatively small as far as I'm concerned, and that statement would not "time out" in any case. It will run until it completes, which should be 3-15 seconds depending on the complexity of the view.
The answer is yes you could make a stored procedure return chunks or ranges of data, but the problem probably lies in the view and retrieving ranges from that is probably not going to improve performance, and making a stored procedure return portions of data from it is probably not the best approach.
To get the first 100 rows
select top 1000 * from mytable
to get the 2nd 1000 rows
select top 1000 * from
( select top 2000 * from mytable order by ID desc )
order by ID ASC
where ID is the unique key in the result set
Make 5 versions of that statement, each retrieving 1 million rows. I can pretty much guarantee that will perform far worse than what you have now though.
Can we see your view? Maybe we can suggest ways to improve the performance.
And yes you can use a view with bcp using the queryout option.
ASKER
Basically, the BCP suggestion led me to discover the SQLBulkCopy commands which are BCP-based.
ASKER