Re: sql comparison between records

-----Original Message-----
From: Joel Fritz <JFritz@sharperimage.com>
To: MIDRANGE-L@midrange. com (E-mail) <MIDRANGE-L@midrange.com>
Date: Friday, July 20, 2001 1:20 PM
Subject: sql comparison between records


>How do you do this in SQL?
>
>I have a file ordered by key value and a timestamp.  There are multiple
>records for each key value. I want to find the difference between time
>stamps of successive records and calculate the average difference for a key
>value group.  I know, it's easy in RPG.  I'm not trying to measure
>programmer productivity. <g>
>
>>Hey. Good question.
>>
>>At first blush:
>>
>>select a.key, avg(a.ts - (select min(b.ts) from theFile b where b.ts >
a.ts
>>and a.key = b.key)) from theFile a group by a.key
>>
>>but I can all but guarantee that that won't work.
>>
>>Maybe with a common table expression or a view....
>>
>>signed, pondering in pennsylvania....   ;-)
>
>
>ok, second blush... it's recusive. that means NO in 400 lingo. In UDB, it
>still probably can't be done because you can't use a column function in the
>recursive subquery....
>
>hhhhmmmmm. Also need some kind of "seq" to locate next.....
>
>still pondering in pennsylvania...
>


And now I'm in search of a V5R1 machine for a couple of minutes.... Al?

Anyway, the theory is this... with 5.1 there is a fetch first row only
clause... maybe a correlated subquery?


select a.key, avg((select b.ts from theFile b where b.ts > a.ts and b.key =
a.key fetch first row only) - a.ts) from theFile a group by a.key


hhhhhhmmmmmmmmmmmmm..........  <seg>


+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---