Multiply rows 1

[blackduck]

Is there a way I can multiply each set of 12 data together eg. Prob for Match1 x Prob for Match2 x Prob Match3.

PosID TimeID MatchID WinID Prob
1 1 1 1 0.521739130434782
1 1 2 1 0.419754953864771
1 1 ..etc 1 0.427902390929258
1 1 12 1 0.461811324143505
2 1 1 1 0.521739130434782
2 1 2 1 0.419754953864771
2 1 ..etc 1 0.427902390929258
2 1 12 2 0.141259934443896

The end result would be something like if you used the aggregate function of sum, but instead I need to multiply.

 

[pjm]

If all your numbers are positive you can use:

select exp (sum (log (ColumnName)) from Table

pjm

 

[blackduck]

It works! Fantastic. Big hug to ya! (and a star of course).

Sorry it took me awhile to get back, I didn't really understand it at first.

For anyone who wants an example, my sql code is:
SELECT tblProbablility.TimeID, tblPossible.PossibleID, Exp(Sum(Log([tblProbablility].[Probablility]))) AS ProductOfProbability
FROM tblPossible LEFT JOIN tblProbablility ON (tblPossible.WinnerID = tblProbablility.WinnerID) AND (tblPossible.MatchID = tblProbablility.MatchID)
GROUP BY tblProbablility.TimeID, tblPossible.PossibleID
HAVING (((tblProbablility.TimeID)<11));

If this looks a little hard. You can use design view, write the query with group/totals and use the sum aggregate until you get it right then instead of the sum column, you just put a new column and type in the field eg.
ProductOfProbability: Exp(Sum(Log([tblProbablility].[Probablility]))) making sure to select Expression where you would have had sum before.

But thanks again to pjm, you saved me heaps.