The techniques the others have cited is basically correct. The only thing to be aware of is that of your data contains floats it might not always work. There are two reasons for this is the floating-point value NaN (not-a-number). Any comparisons involving it will come out False. Hence:
(1 < NaN) = false
(1 > NaN )= false
(NaN = NaN) = false
(NaN <> NaN) = false
The second problem is rounding error. As someone on the forum found out recently:
(4.7 < 4.7) = false
((4.6 +.1) < 4.7) =
TRUEIf the cluster is all data that is user input you'll probibly get away with it, but if there are calculations involved in deriving the values you are on potentially very thin ice. Food for thought...
Mike...