Creating Subsets

[Riot]

So I am tasked with creating subsets for my data. We have a couple hundred columns. So it is basically Create subset "Hypertension" is variable Clinic01 == 1 but there are many different subsets I have to create. In my head, I imagine that I can get there by creating a bunch of Filter Tools, but I have about a hundred subsets to create, so i was wondering if there exist a faster and more efficent way to create subsets. I have lookedbut have been unable to find them. Thank you. 

[MarqueeCrew]

Robert,

If you are outputting the subsets, you can take name from a field and append it to the filename. You can keep the field in the data and create hundreds of output files from just a single output tool.

Thanks,

Mark

[Riot]

I am not quite following you good sir. 

For one section, my "WholeClean" file I have  :

Hypertension <- subset(WholeClean, clinic01 == 1)

 Hypertension_Limited <- subset(WholeClean, clinic02 == 1)

Angina <- subset(WholeClean, clinic03 == 1)

 CoronaryArteryDis <- subset(WholeClean, clinic05 == 1)

 MyocardialInf <- subset(WholeClean, clinic09 == 1)

 Arthritis <- subset(WholeClean, clinic19 == 1)

Arthritis_Limited <- subset(WholeClean, clinic20 == 1)

 Migraines <- subset(WholeClean, clinic21 == 1)

Migraines_Limited <- subset(WholeClean, clinic22 == 1)

 Asthma <- subset(WholeClean, clinic23 == 1)

Asthma_Limited <- subset(WholeClean, clinic24 == 1)

DiabetesSym <- subset(WholeClean, clinic31 == 1)

 Depression <- subset(WholeClean, clinic44 == 1)

 Depression_Limited <- subset(WholeClean, clinic45 == 1)

 Anxiety <- subset(WholeClean, clinic49 == 1)

 Anxiety_Limited <- subset(WholeClean, clinic50 == 1)

So the first part is the name of the subset I am creating, WholeClean is the name of my file, clinic## is the column I am testing. 

[MarqueeCrew]

If @jjdunkerly doesn't get you an answer by tonight, I will work with you tomorrow am. Sorry, I'm at an Alteryx roadshow without my laptop.

[MarqueeCrew]

Correction @jdunkerley79

[Riot]

Thank you, I appreciate it. 

[jdunkerley79]

Happy to have a go @MarqueeCrew,

Could you send a sampe set of data to work with (by all means DM if you need to)?

[jdunkerley79 · Accepted solution]

Ok so here is a first go.

I take the data and transpose it so looks like:

Filter this where the value is 1. I am working with a numeric field but this can be changed to cope as needed.

After that I join to a table mapping clinic to subset name

This gives a set of subsets each case belongs to

Next join back to the original data to add a subset field and duplicate the rows for those cases in multiple subsets

Finally as @MarqueeCrew mentioned use an Output tool to save to a set of files append the subset name to the file name to create a set of files one for each subset

Sample attached with some fake data. Will Write to C:\Temp when run

[Riot]

That looks good. I am going to try all f my data like that and see how it works out. Thank you so much. Im still a little unclear with the output thing, but I will try and figure it out. 

[jdunkerley79]

The output tool has a config section that controls saving to different files:

Very useful feature.