MMDDYY10 date format question

[RobF]

I'm creating a simple toy dataset containing date variables with the MMDDYY10. format:

data test;

input id $ ed_from MMDDYY10. ed_to MMDDYY10. ip_from MMDDYY10. ip_to MMDDYY10.;

format ed_from MMDDYY10. ed_to MMDDYY10. ip_from MMDDYY10. ip_to MMDDYY10.;

datalines;

CT0001 05/22/2011 05/22/2011 05/22/2011 05/30/2011

CT0001 06/22/2011 06/22/2011 06/22/2011 06/30/2011

CT0002 07/01/2012 07/02/2012 07/01/2012 07/06/2012

CT0002 07/10/2012 07/10/2012 07/20/2012 07/25/2012

;

run;

However I receive "Invalid data..." error messages when I attempt to run the above code.

Any help identifying what I'm doing wrong is much appreciated!

Thanks,

Robert

[DBailey · Accepted solution]

try this:

input id $ ed_from:MMDDYY10. ed_to:MMDDYY10. ip_from:MMDDYY10. ip_to:MMDDYY10.;

[RobF]

Fantastic, worked beautifully. I wasn't aware we should use colons in the input statement, this is good to know.

Thanks DBailey!

[PGStats]

You can use either formatted input with a field size that correspond with the TOTAL column spacing:

input id $ (ed_from ed_to ip_from ip_to) (MMDDYY11.);

or use pointer controls to move the line pointer to the beginning of date fields:

input id $ (ed_from ed_to ip_from ip_to) (+1 MMDDYY10.);

or modified list input with an informat that correspond with the size of the text tokens irrespective of the spacing between them:

input id $ (ed_from ed_to ip_from ip_to) (:MMDDYY10.);

PG

Message was edited by: PG Added an alternative.

[RobF]

Thank you, this is very helpful!