Sql question

bfuchs
bfuchs used Ask the Experts™
on
Hi Experts,
I have the following table
TmpPatBillingCodes:

TmpPatBillingCodesID Number
EffectiveFrom Date
End Date

FYI-The first Column is not unique, its linked to another record of this table.
there are multiple records with same TmpPatBillingCodesID but different Dates.
Now I need to get a count of TmpPatBillingCodesID  grouped by datediff from effectivefrom-end 30 days, 60 days, 90 days and 90+
One more thing, All records containing the same TmpPatBillingCodesID are counted as 1 regarding the dates, meaning I need to get the date difference between min(EffectiveFrom) and Max(End) grouped by TmpPatBillingCodesID.
I would appreciate your help with this.
I am posting the sql that I have tried in Access, however it gives me an error that aggregate function in group by clause..
SELECT Count(TmpPatBillingCodes.TmpPatBillingCodesID) AS CountOfTmpPatBillingCodesID, Switch(DateDiff("d",Min([effectivefrom]),Max([end]))<31,"1-30",DateDiff("d",Min([effectivefrom]),Max([end])) Between 31 And 60,"31-60",DateDiff("d",Min([effectivefrom]),Max([end])) Between 61 And 90,"61-90",DateDiff("d",Min([effectivefrom]),Max([end])) Between 91 And 180,"91-180",DateDiff("d",Min([effectivefrom]),Max([end])) Between 181 And 365,"181-365",DateDiff("d",Min([effectivefrom]),Max([end]))>365,"365+") AS Expr2
FROM TmpPatBillingCodes
WHERE (((TmpPatBillingCodes.End) Between #1/1/2012# And #12/31/2012#))
GROUP BY Switch(DateDiff("d",Min([effectivefrom]),Max([end]))<31,"1-30",DateDiff("d",Min([effectivefrom]),Max([end])) Between 31 And 60,"31-60",DateDiff("d",Min([effectivefrom]),Max([end])) Between 61 And 90,"61-90",DateDiff("d",Min([effectivefrom]),Max([end])) Between 91 And 180,"91-180",DateDiff("d",Min([effectivefrom]),Max([end])) Between 181 And 365,"181-365",DateDiff("d",Min([effectivefrom]),Max([end]))>365,"365+");

Open in new window

 

I know it would be easier to get this done in separate queries, however since this is meant for a user to run it temporarily, I would prefer not to have to save queries therefore my goal here is to have this all in one query, so user can just copy and paste.

Thanks
Comment
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
Most Valuable Expert 2015
Distinguished Expert 2018
Commented:
You should study the weird and little known function Partition() here:

http://msdn.microsoft.com/en-us/library/office/gg264461(v=office.15).aspx

It might come in handy.

/gustav
@Gustav
this seems a little to complicate (or weird as you called) for me to understand and implement..
however I managed to get this work with the following
select count(TmpPatBillingCodesID) ,expr2
from
(SELECT TmpPatBillingCodes.TmpPatBillingCodesID, Switch(DateDiff("d",Min([effectivefrom]),Max([end]))<31,"1-30",DateDiff("d",Min([effectivefrom]),Max([end])) Between 31 And 60,"31-60",DateDiff("d",Min([effectivefrom]),Max([end])) Between 61 And 90,"61-90",DateDiff("d",Min([effectivefrom]),Max([end])) Between 91 And 180,"91-180",DateDiff("d",Min([effectivefrom]),Max([end])) Between 181 And 365,"181-365",DateDiff("d",Min([effectivefrom]),Max([end]))>365,"365+") AS Expr2
FROM TmpPatBillingCodes
WHERE (((TmpPatBillingCodes.End) Between #1/1/2012# And #12/31/2012#))
GROUP BY TmpPatBillingCodes.TmpPatBillingCodesID) as t
group by expr2

Open in new window


Thanks,
Ben
Most Valuable Expert 2015
Distinguished Expert 2018

Commented:
OK, but basically it does what your subquery does and probably a bit faster.

Partition is not difficult to use, it's just that you do it so rarely that you forget how. I think I have use it twice - and I have used Access since version 1.0.

/gustav
thanks
Most Valuable Expert 2015
Distinguished Expert 2018

Commented:
You are welcome!

/gustav

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial