Reading raw file data with different date type in different records

[sunilreddy]

I've a raw file with different date values in different records.

eX:

Datalines;

08DEC2015

12112014

201113

;

How to read different date types from raw file.

[Reeza]

You can try ANYDTDTE or you can bring it in as character and then try and parse it out using some logic if/then statements.


data want;
format date date9.;
input date anydtdte.;
cards;
08DEC2015
12112014
201113
;
run;

proc print data=want;
run;

[ballardw]

Don't be surprised if you get values you don't expect.

Your example of 12112014 will depend on your national language setting for whether that is supposed to be MM/DD/YY or DD/MM/YY.

And 201113 is likely to have problems if it is supposed to be YYYYMMDD as it could be interpreted as 2011 Jan 3 or 20 Nov 2013

[Reeza]

Here's an example of a thorough process...it's a bit of work. I've seen some cleaner regex solutions, but I can't seem to find any right now.

[RW9]

Personally I wouldn't risk trying to work with that data.  Go back to your data import agreement - you have on of those right? - and check what it states in there.  If you dont have one of those, make one and get the vendor to agree to it.  They should be providing a robust, structured dataset, which they currently are not.  For instance, what does 12112014, mean?  Is it december or november.  As with everything programmatic - specify, specify, specify, then code.  Programming is in fact 99% irrelevant, specification and documentation is the key.