Date formats

[india2016]

03JAN2011:00:00:00.000

which format should I use to read this date ?

[art297]

Depends upon what you want. That is a datetime, not a date. If you want to capture all three you could use something like:

data have;
  input dt datetime22.;
  format date date9.;
  format time time5.;
  date=datepart(dt);
  time=timepart(dt);
  cards;
03JAN2011:00:00:00.000
;

Art, CEO, AnalystFinder.com

[ChrisBrooks]

You need an informat to read data not a format - try datetime22. (Untested)

[ballardw]

If all of the values have the same "time", especially of :00:00:00.000 I would use a DATE9. informat to read the date and discard the time part.

Don't forget to set a display format so humans can see the date in a useable form.

[Reeza]

When in doubt try ANYDTDTM for datetime and ANYDTDTE for dates.

[ChrisBrooks]

The only note of caution I'd sound about ANYDTDTM and ANYDTDTE is that they are dependant on the correct Locale being set. I learned that the hard way a few years ago when working on a server in the U.K. where unknown to me the Locale had been set to US which led to some correct dates, some incorrect dates and some errors. It puzzled me for quite a while before I realised what was going on.....

[india2016]

Thanks, everyone

.

I used Datetime22.3 format and it successfully reads that observation.