on Mon, May 28, 2001 at 01:04:07AM +0000, Hugues Richard (hugues@PANTERA.RI.MGH.MCGILL.CA) wrote: > Hello , > > I have a file in a julian format (yyddd) similar to the following > example . I am sure it's obvious , but what I am doing wrong ? > > DATA un ; > INPUT matb $ @4 deces julian5. ; > CARDS ; > m1 95120 > m2 00000 > ; > PROC PRINT ; > thanks a lot, > hugues > > > log: > NOTE: Invalid data for deces in line 882 4-9. > RULE:----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8----+----9-- > > 882 m2 00000 > matb=m2 deces=. _ERROR_=1 _N_=2 > NOTE: The data set WORK.UN has 2 observations and 2 variables. > NOTE: DATA statement used: > real time 0.03 seconds > cpu time 0.00 seconds

I believe that this is indicating that '00000' is considered an invalid julian date. There is no 'day 0' of a year.

You may want to verify this with some data step code testing various Julian values. The SAS online documentation indicates that valid day values are in the range 1-366.

I believe you'll have to add some conditional processing to test your data before assigning it to a julian date, e.g.: (untested code, and I'm getting rusty):

drop _juldate; input @4 _juldate $4.; if 1 le input( substr( _juldate, 3, 3 )) le 366 then deces = input( _juldate, julian5. ); else deces = . ;

Cheers.

-- Karsten M. Self <kmself@ix.netcom.com> http://kmself.home.netcom.com/ What part of "Gestalt" don't you understand? There is no K5 cabal http://gestalt-system.sourceforge.net/ http://www.kuro5hin.org Disclaimer: http://www.goldmark.org/jeff/stupid-disclaimers/