basically, its a survey dataset that have a score of 1 to 10..
every quarter, we need to exclude 1 lowest and 1 highest.
currently we are doing the manual exclusion of the lowest and highest, we're wondering if we can do it in beastmode
I don't think you'd be able to do that in beastmode unless I'm not understanding something. You'd have to do it in ETL.
Say MySQL, If the surveys come over with some ID you could do something like this...check my syntax and logic...but theory should be good. My thought here is you may have multiple surveys with the same lowest and highest value.
transform 1 - determine the value
select min(score) as `min score`, max(score) as `max score` from surveylist
transform 2 - find low
select top 1 `survey ID`
from surveylist as l, transform1 t
where l.score = t.minscore
transform 3 - find high
where l.score = t.highscore
select * from surveylist
where `survey ID` not in (select `survey ID` from transform2) -keeps 1 survey that had the lowest out
and `survey ID` not in (select `survey ID` from transform3)-keeps 1 survey that had the highest out
Thanks
Matt
