topic Re: Return value in one column based on location of a value in another column in Discussions

Return value in one column based on location of a value in another column

MarkDayton — Wed, 11 Dec 2019 17:02:16 GMT

I have a dataset where devices are identified by lot, wafer, and die; and each part may have multiple tests. I can search the test date column to find when the last test was run on each die, I then need to know from the pass_fail column what the status of the last test was, something like:

:PASS_FAIL[ row of Col Maximum(:TEST DATE, :LOTID, :WAFER, :DIE)]

I am hoping for a formula that I can use in Graph Build to generate pass/fail wafer maps based on the state of the last test for each die.

Re: Return value in one column based on location of a value in another column

MarkDayton — Wed, 11 Dec 2019 17:10:30 GMT

Nevermind, I can add a column to calculate the column maximum, and then select where the test date matches the column maximum test date for each die, and then create a subset of the data from the selected rows.

Re: Return value in one column based on location of a value in another column

MarkDayton — Thu, 12 Dec 2019 14:46:30 GMT

dt << Select Where( :Test Date == Col Maximum(:TEST DATE, :LOTID, :WAFER, :DIE))

Re: Return value in one column based on location of a value in another column

MarkDayton — Thu, 12 Dec 2019 14:47:11 GMT

Better Solution:

In Graph Builder create a transform formula:

:TEST DATE == Col Maximum(:TEST DATE, :LOTID, :WAFER, :DIE)

Then add a Data Filter to select where the transform = 1, no subset required.