Thank you Arthur, I appreciate your help.

Mike

On Mon, 11 Jan 2010 16:09:26 -0500, Arthur Tabachneck <art297@NETSCAPE.NET> wrote:

>Mike, > >Would something like the following suffice?: > >data have; > informat sales_date date7.; > format sales_date date7.; > input PERSON_ID SALES_DATE UNIT_ID; > cards; > 77 01SEP06 2100 > 77 01SEP06 3557 > 77 05NOV06 1454 > 77 05NOV06 3765 > 77 13NOV06 1345 > 77 13NOV06 1456 >; > >proc transpose data=have out=want (drop=_:); > var unit_id; > by person_id sales_date; >run; > >data want (drop=col:); > set want; > unit_ids=catx('-',of col:); >run; > >HTH, >Art >------- >On Mon, 11 Jan 2010 15:28:05 -0500, Mike Green <mike_green123@YAHOO.COM> >wrote: > >>Hi, >> >>I would appreciate it if someone could help me with problem. >> >>I need to add a variable to the dataset that is a concatenation of the >>UNIT_ID variable by SALES_DATE. Example - for 01SEP06 I need a new unique >>id of 2100-3557. >> >> Obs PERSON_ID SALES_DATE UNIT_ID >> >> 137 77 01SEP06 2100 >> 138 77 01SEP06 3557 >> 139 77 05NOV06 1454 >> 140 77 05NOV06 3765 >> 141 77 13NOV06 1345 >> 142 77 13NOV06 1456 >> >>Thanks, >> >>Mike