topic Selecting the highest GDG in a dataset name in SAS Programming

Selecting the highest GDG in a dataset name

TingSern — Mon, 29 Jan 2018 12:12:25 GMT

I have the following observations ....

DATA INPUT;

DSNAME -

GPSP.BKUP.EXPIRED.TRANS.LOG.G3821V00
GPSP.BKUP.EXPIRED.TRANS.LOG.G3822V00
GPSP.BKUP.EXPIRED.TRANS.LOG.G3823V00
GPSP.BKUP.EXPIRED.TRANS.LOG.G3824V00
GPSP.BKUP.EXPIRED.TRANS.LOG.G3825V00
GPSP.BKUP.EXPIRED.TRANS.LOG.G3826V00
GPSP.BKUP.EXPIRED.TRANS.LOG.G3827V00
GPSP.BKUP.EXPIRED.TRANS.LOG.G3828V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0033V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0034V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0035V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0036V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0037V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0038V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0039V00

I like to select (in this example) - only the following entries -

GPSP.BKUP.EXPIRED.TRANS.LOG.G3828V00

GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0039V00

I tried the following code -

DATA WANTED;
SET INPUT;
BY DSNAME;
IF LAST.DSNAME;
RUN;

It does NOT work. Thank you.,

Re: Selecting the highest GDG in a dataset name

Kurt_Bremser — Mon, 29 Jan 2018 12:31:07 GMT

It's because your dsname is always unique, and therefore each by group contains only 1 observation.

I guess you would want this:

data input;
input dsname $50.;
dsname_short = substr(dsname,1,find(dsname,'.',-50)-1);
cards;
GPSP.BKUP.EXPIRED.TRANS.LOG.G3821V00      
GPSP.BKUP.EXPIRED.TRANS.LOG.G3822V00      
GPSP.BKUP.EXPIRED.TRANS.LOG.G3823V00      
GPSP.BKUP.EXPIRED.TRANS.LOG.G3824V00      
GPSP.BKUP.EXPIRED.TRANS.LOG.G3825V00      
GPSP.BKUP.EXPIRED.TRANS.LOG.G3826V00      
GPSP.BKUP.EXPIRED.TRANS.LOG.G3827V00      
GPSP.BKUP.EXPIRED.TRANS.LOG.G3828V00      
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0033V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0034V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0035V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0036V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0037V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0038V00
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0039V00
;
run;

data want;
set input;
by dsname_short;
if last.dsname_short;
run;

proc print data=want noobs;
run;

Result:

                  dsname                                dsname_short

GPSP.BKUP.EXPIRED.TRANS.LOG.G3828V00          GPSP.BKUP.EXPIRED.TRANS.LOG      
GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY.G0039V00    GPSP.BKUP.EXTR.CD.GPPFINTF.WR.MLY

Re: Selecting the highest GDG in a dataset name

TingSern — Tue, 30 Jan 2018 05:22:41 GMT

Thank you. It works like a charm.