Grouping on a numeric field: Doesn't seem to work in this query

[PeteCresswell]

I'm trying to smoke out duplicate return records for Fund/ReturnDates
in which the ReturnRate differs.

We are loading a DB using data from several vendors and if a given
fund's return for a given date differs among vendors, we want to know
about it.

-----------------------------
SELECT ttblReturnStage_Monthly.FundID,
ttblReturnStage_Monthly.ReturnDate,
ttblReturnStage_Monthly.ReturnRate

FROM ttblReturnStage_Monthly

GROUP BY
ttblReturnStage_Monthly.FundID,
ttblReturnStage_Monthly.ReturnDate,
ttblReturnStage_Monthly.ReturnRate

HAVING (((Count(ttblReturnStage_Monthly.RecordID))>1));
-----------------------------

ttblReturnStage_Monthly.ReturnRate is a Double with six decimal places.

The actual values are mostly only four decimal places.

Can anybody see what I'm doing wrong that this query thinks many
records with the same FundID/ReturnDate/ReturnRate are dupes?

[PeteCresswell]

I guess I should add that ttblReturnStage_Monthly.RecordID is an
AutoNumber PK.

[PeteCresswell]

> GROUP BY

ttblReturnStage_Monthly.FundID*,
ttblReturnStage_Monthly.Return*Date,
ttblReturnStage_Monthly.Return*Rate
Oops! RCI strikes once again. I see the error of my ways.

SHB:
--------------------------------- GROUP BY
ttblReturnStage_Monthly.FundID*,
ttblReturnStage_Monthly.Return*Date

(+ a few more queries to ID the rates that don't agree...)
---------------------------------