Percentile.vi should propage "NaN" values

[ec01]

Hello,

now the percentile VI calculate a table with NaN-values as the NaN-values are numbers higher than +Inf. This is incorrect.

The percentile VI should propage to result NaN if one of the elements of the table is NaN.

(An other solution is not consider NaN values and return as result the value greater than p percent of the data values in the array, but it would create incoherences with all other VI which propage the "NaN value")

Here a screenshot of the a bad response of this VI.

[altenbach]

In what universe are NaN higher than +Inf? You can look at the code and see that it basically does a sorting followed by threshold array. Sorting places NaN at the end. I agree that the presence of NaN should cause a NaN output. Similarly, if you wire NaN to the "p" input, it currently gives a result that is not NaN. Not good!

[AristosQueue (NI)]

I filed this (including Altenbach's comment) as CAR 529474. This corrective action request may be fixed in a future version of LabVIEW as priority permits. Thank you for reporting the issue.

[Darren]

Moved to CAR database: CAR 529474