So, i have a code given to me by a previous DBA for how partitioning is setup currently in our env, I am confused by the use of multiple files on a filegrp in the code, so here it goes:
Looks like partitioning is setup by year, The PS is mapped to PF by year and the data from a year say 2013 goes to a FG say named FG_2013, all good till now. What is confusing me is that the FG_2013 contains 12 files (01.ndf and so on till 12.ndf) supposedly to mean per month, but SQL doesn't know 01 means data from Jan right? and then the code marks 01-03 as quarters and so on for the 4 quarters, to me it looks like the code is thinking that SQL knows 01 is Jan , 12 is Dec but it doesn't it just keeps data in the 12 files randomly (right?).
Simpler would be to keep 1 file per FG - meaning all 2013 data goes to a single file in FG_2013 or to involve PF-PS code using months/quarters. I am just confused, let me know if what I am thinking is right (SQL has no way to know 01.ndf file means Jan - as nothing on the code defines that & thus data is just kept yearly on the fg randomly and 01.ndf is not Jan data only?). Thanks.
D