Should I copy the entire table or only the columns | SQL Server Performance Forums

SQL Server Performance Forum – Threads Archive

Should I copy the entire table or only the columns

I need to copy data from one database to another on a daily basis.
The source table has around 500 columns and is about 2GB. I only need about 100 columns on the destination. Space is not really an issue, but performance is. Will it be moer efficient to copy the entire table across every day, or would it be faster to only copy across the columns I need? What is the best way to do this? Snapshot replication, stored procs or DTS??
Thanks
Kate
Using DTS task you can setup a job and schedule it, once for all you can select the required columns. If its not a problem then simply export the whole table to the destination table and run query with required columns. I believe there would be no difference in performance exporting whole table and selective columns. _________
Satya SKJ

Copying a few columns should definitely be faster (few bytes).
Under the same conditions DTS would be easier to configure and schedule. By the way, is the copying going to occur within the same box, on the LAN or …. ?
I believe it may not be much difference between 2 jobs. _________
Satya SKJ

OK, for now 400MB vs 2GB may not reflect a big difference on the same box or on a fast Ethernet LAN. That’s why I was wondering whether a WAN was involved…… What happens later when then data is 20GB+ or is this the proverbial "crossing of bridges when we come to them"?
Hi Just go through with DTC and select only required columns as per ur requirement Rushendra
Rushi Are you referring about DTS or DTC??? _________
Satya SKJ

…In either case, a scheduled "DTS" task has already been suggested by Satya. If Rushi is referring to MS "DTC", I would not find any place for it in this scenario – Kateperkins just wanted to copy some data across.
NHO
Sorry. It is DTS only not DTC.
quote:Originally posted by satya Rushi Are you referring about DTS or DTC??? _________
Satya SKJ

Rushendra
]]>