I think I may be overthinking this problem.
What I have is a table that consists of membership data including visit dates. For each visit date, I need to determine the number of days since the last visit.
I'm trying a combination of CTEs and partitions, but it is simply not coalescing in my head.
Following is a sample set of data;
KeyID MemberID AdmitDate Last_Admit_Days
1 2601 2014-04-17
2 2001 2014-05-04
3 1001 2014-05-15
4 8001 2014-05-28
5 2601 2014-05-23
6 2601 2014-05-28
7 9301 2014-03-21
8 9301 2014-05-15
9 9301 2014-05-21
10 8601 2014-05-26
11 3601 2014-04-13
12 3601 2014-05-05
I tried putting that in as a code block, but it ended up putting everything on one line.
Anyway, Member # 2601 has three admit dates; 2014-04-17, 2014-05-23 and 2014-05-28. So, the number in the Last_Admit_Days column for those records, should be the DATEDIFF in days between the AdmitDate and the AdmitDate of the immediately prior AdmitDate for #2601, like this;
KeyID MemberID AdmitDate Last_Admit_Days
1 2601 2014-04-17
0
5 2601 2014-05-23
36
6 2601 2014-05-28
5
I am stumped. Maybe the weekend will refresh me to take a better stab at it, but it's confounding me at the moment and I thought I'd ask here to see if anyone had a quick idea off the top of their heads.
I truly appreciate any assistance.
Thanx!