Randomly select rows within a given column value

[#1]

Hello everyone, 

First of all, I'm an occasional JMP user (v16.2) with little JSL experience and limited statistical knowledge. I couldn't directly find an answer on the community, I hope I'm not asking something which has already been answered. That being said, I've been trying to figure out how to do the question below 'manually' - using standard JMP functionalities. I couldn't find any simple way, so I hope somebody can give me a hint using JSL.

For a particular application request I would like to subset a large data table. More specifically, I would like to subset the table by randomly taking out a number of rows, for a specific column value. For example, I'd like to transfer the table attached to a new table, where I can obtain 3 (randomly chosen) values of column "Numerics" for each column value of column "condition". 

Apart from the JSL solution, is there maybe a way to do this using any of the statistical platforms?

Thanks for the help!

[#2 · Accepted solution]

Hi @Nico_M,

Welcome in the Community !

If I understood correctly your need, you should be able to do this using the "Subset" platform in menu "Tables" :

1. Click on menu "Tables", then "Subset" ; a new window opens,
2. Specify "random - sample size" as 3 (or any number of rows you want per condition) and select stratify and choose the column "Condition" (as shown in the screenshot) :
   
3. Click on "Ok" and you'll get a new subset table :)

Here is the JSL code on your example if you want to directly use it :

Data Table( "JMP replicates - community" ) << Subset(
	Sample Size( 3 ),
	Selected columns only( 0 ),
	Stratify( :Condition )
)

I also send you back your datatable with the Subset table saved as a new script.

Hope this answer will help you, 

[#3]

Thanks! I completely missed the 'stratify' option, it was indeed very simple ;).