topic Re: How to create an incremental grouping variable based on lagged values? in SAS Programming

How to create an incremental grouping variable based on lagged values?

Ani7 — Wed, 11 May 2022 15:52:21 GMT

I have a table of values with the relevant column containing information on which rows need to be combined in my final table that looks like this:

to_combine
1
.
1
1
.
1
1
1
1
.
1
.
.

I would like to create a new variable that creates a new group for each of these "groups of 1s" that would look like this

to_combine	group
1	1
.	.
1	2
1	2
.	.
1	3
1	3
1	3
1	3
.	.
1	4
.	.
.	.

How would I go about creating this?

I tried creating an "obs" variable and setting it so that if to_combine = 1 then obs = lag(obs). However, that seems to set obs to the last row that had to_combine = 1 instead of the immediately preceding row.

Re: How to create an incremental grouping variable based on lagged values?

PaigeMiller — Wed, 11 May 2022 16:27:43 GMT

You don't state explicitly what the rule is for creating a new group. It always helps to be clear and not expect us to figure it out (sometimes we get it wrong). So is the separator for a new group one missing value? What if there are multiple missing values in a row?

Re: How to create an incremental grouping variable based on lagged values?

Ani7 — Wed, 11 May 2022 16:54:26 GMT

That's my bad. Essentially all consecutive 1s in the to_combine column should have a unique group number (missing rows are indicative of a new group number that needs to be started).

Re: How to create an incremental grouping variable based on lagged values?

PaigeMiller — Wed, 11 May 2022 17:23:39 GMT

Thanks

data want;
    set have;
    if _n_=1 then newgroup=1;
    if missing(to_combine) then newgroup+1;
    if not missing(to_combine) then group=newgroup;
    else group=.;
    drop newgroup;
run;