Can we use Recode to classify values as missing values

[#1]

1)I want to classify negative values as missing values. Can we use Recode to  accomplish this. If yes, how can we do it?  I can create a new formula column while using Recode. I see it as a two step process. I was wondering if this could be accomplished in a single step.

2) If we use data filters, can recode be used in the subset of the filtered data with out affecting the rest.

[#2 · Accepted solution]

The easiest way to do this is to create a new column and then just use the following formula

If( :your variable < 0, ., :your variable )

[#3]

Nelson,

     Thanks. I was able to implement your solution. I have a Q for you. Here is another way I tried before I posted the Q: When I tried to select rows having negative values in Recode and then asking it to replace with the dot, then JMP warned that it is going to change the column data type into character.(I copied the original column into a new one so that I do not alter original values and performed my changes in the new column).But when I used your formula, it did not. I think we are trying to accomplish the same thing in different ways.  So what is the difference? Looks like the dot when used  in JSL takes the definition of the data type of that column-Is that right? If not, could you please explain.

[#4]

I have not been able to replicate your findings.  I checked out versions JMP 13-JMP 16 and did not observe your findings

[#5]

Dot is not the value for a missing observation. It is JMP's way of representing it in the data table or a script for your benefit. Just leave the value blank in Recode. In other words replace the target with nothing.

If you enter "." then JMP thinks it is a character value. (Not a missing character value, either!)

The only time the period character is used is in a script, such as an assignment. There has to be a value in an assignment:

Assign( left, right ) // assign right value to left variable

left = right