topic how to sort and compare columns when data is not identical in Discussions

how to sort and compare columns when data is not identical

Ohad_s — Sat, 10 Jun 2023 23:59:05 GMT

I have data from two separate runs.

data include: X,Y,Z

I want to comapre results based on X,Y coordinates (Z is the results)

the problem: for each run i get different coordinates. well, not completely different but not identical. within a range of +- 10.

how can i sort the data and compare the results (it has to be X and Y, not one of them only).

see table for example, (I've arranged the data in that order just so it will be easy to see)

working with JMP 15

Re: how to sort and compare columns when data is not identical

Thierry_S — Thu, 29 Dec 2022 00:44:43 GMT

Hi,

It is challenging to address your question because "comparing" data could mean different things. Anyway, if you want to sort the Z results by a common X and Y, you may want to create an aggregated value for Xs (e.g., Mean (:X1, :X2)) and for Y (e.g., Mean (:Y1, :Y2), as long as the intervals for your Xs and Ys do not overlap.

If you are interested in testing the effects of X and Y on Z, you could stack your columns Stack > X1, X2, Y1, Y2, Z1, Z2 > multiple series = 3; Contiguous.

Are these two ideas of any help?

Please feel free to reply to this post with more details.

Best

Re: how to sort and compare columns when data is not identical

Ohad_s — Wed, 04 Jan 2023 09:21:54 GMT

actually what I did is sum X1Y1 and X2Y2 and deduct the results in abs values.

that way i could set a zone of search for each results and compare Z values that are within the zone of search

Re: how to sort and compare columns when data is not identical

Jed_Campbell — Wed, 04 Jan 2023 15:21:07 GMT

Another way might be to round each of your X and Y values to the nearest 10 units, then concatenate them. I did this in 3 new formula columns in the attached data table, but it could easily be done in just one column. This would reduce the 2 dimensions to just one, after which this new variable could be the basis for whichever comparisons you're planning.