hey i am working on a LARGE analysis services/datamart project where i need to group by different networks of ip addresses.
i have more than 64,000 IP addresses (especially for the destination fields)
i need to:
1) store these the most efficiently (smallest). isnt integer the smallest way to do this?
2) convert these (to the destination format the most efficiently)
3) roll these up into different networks (ie i think that i want a table where i store:
my preferences are to not use any DTS/vba for this (because i am using bcp/bulk insert)
and the UDF that I have previously used seems WWWWWAAAAAAAAYYYYYYY tooo slow.
i have seen people use an extended stored proc to do this (written in C++ i assume)-- is this utterly the fastest way to do this?
i am talking about 300million new records every day.