Harry,

All four methods produce the same results but, being a strong believer in parsimony, I have to vote for Howard's code.

Art --------- On Tue, 29 Apr 2008 19:55:14 -0400, Howard Schreier <hs AT dc-sug DOT org> <schreier.junk.mail@GMAIL.COM> wrote:

>On Tue, 29 Apr 2008 19:09:06 -0400, Arthur Tabachneck <art297@NETSCAPE.NET> >wrote: > >>Harry, >> >>Who am I to question data_null_? >> >>Regardless, I got it to work with input--input. E.g., >> >>data have; >> input num_dates; >> cards; >>20060607 >>20050822 >>20080429 >>; >> >>data want; >> set have; >> format date date9.; >> date=input(input(num_dates,$12.),yymmdd8.); > >Or just > > date=input(num_dates,yymmdd12.); > >>run; >> >>HTH, >>Art > >Also, there is an all-numeric approach: use numeric functions and operators >to extract the year, month, and date components, then use the MDY function >to construct the SAS date. > >>------- >>On Tue, 29 Apr 2008 18:50:47 -0400, Hari Nath <hari_s_nath@YAHOO.COM> >>wrote: >> >>>Hi all, >>> >>>I have a sas dataset which has numeric dates in the form 20060822 >>>all my proc contents says is that >>> >>> # Variable Type Len Label >>> >>>137 START_DT1 Num 8 START_DT1 >>> >>>I want to find a way to read them and format it to mmddyy8. >>> >>>am not successful so far....can someone guide me here..... >>> >>>Many thanks >>>hari.