How to create a group-ID for interlinked ID numbers?

[KarenGeurts]

I have a dataset with 2 columns of ID numbers (id1 id2). Each row establishes a link between 2 ID numbers There are several many-to-many links, and some ID numbers appear in both columns. I want to create a "group-ID" for all IDs which are interlinked with each other. How can I do this?

In the example below, I want to create 2 group IDs: one for ID numbers a b c d, and another for ID numbers e f.

data have;

input (id1 id2) (:$1.);

cards;

a b

a c

b d

e f

;

Thanks for your help.

Karen

[Reeza]

If you have SAS/OR licensed then look at PROC BOM.

If not, there's a good solution posted on here earlier that some searching may help find.

[KarenGeurts]

Thank for your answer, but I do not have SAS/OR.

I used the macro SubGraphs as suggested by PG, which works very well

Karen

[Reeza]

I don't have SAS/OR anymore either, so I'm glad to see PG's macro as well!

[PGStats · Accepted solution]

This can be done with my SubGraphs macro. See

How to find all connected components in a graph

PG

[KarenGeurts]

It works wonderfully.

Thank you very much

Karen

[Reeza]

@ChrisHemedinger

Where would this document now exist?

[ChrisHemedinger]

@Reeza - I fixed the link to point to the proper article.