Indicate the first record in a data set

[Paul_NYS]

Hi

In processing a data step, is there a way to indicate the first record in an overall data set? I know using 'first' works to indicate the first grouping of a certain combination of variables, but I am wondering about the entire data set.

Paul

[PaigeMiller · Accepted solution]

The predefined variable _n_ will have the value 1 for the first record in a data set

[Paul_NYS]

Thanks Paige.

Paul

[PGStats]

If your datastep involves reading a single dataset then testing _N_=1 will signal the first datastep iteration. Otherwise, using the INDSNAME=myVariable option in the SET statement will give you a dataset name to compare with the previous iteration (IF myVariable ne lag(myVariable)) to detect the fiirst observation from each new dataset, including the first one.

PG

[Paul_NYS]

Thanks PG. I have one data set, but that is good to know for the future.

Paul