topic Re: compress specific sequence of chars in SAS Procedures

compress specific sequence of chars

Jsendzik — Thu, 01 May 2014 14:09:18 GMT

Hey,

I was give a bunch of data sets that have been reworked to include a prefix on the variable name. I have a loop written to remove the prefix but the problem I'm encountering is the compress function is removing all characters within the prefix from the var name. I'd like it to remove the specific sequence of the letters and not all the letters. Is it possible to do this?

The prefix in all of these example is "INF_"

Input	Current_Result	Desired_Result
INF_FINANCE	ACE	FINANCE
INF_INFO	O	INFO
INF_CAST	CAST	CAST

Not all of the variables in the data set have the prefix, that is why I've attempted to do this using the compress function.

Thanks!

Re: compress specific sequence of chars

Jsendzik — Thu, 01 May 2014 14:16:59 GMT

Found out how to do this using the TRANWRD function.

Re: compress specific sequence of chars

RW9 — Thu, 01 May 2014 14:25:32 GMT

Yep, just be careful though as if you have INF_INFO and you do a tranwrd(...,'INF','') then you would end up with _O. If you know the prefix is INF_ first three characters, it could be beneficial to use substr, e.g. if substr(...,1,4)='INF_' then ...=substr(...,4);

Re: compress specific sequence of chars

Tom — Thu, 01 May 2014 16:39:59 GMT

Use the TRANSTRN() function to replace the string INF_ with nothing.

desired_result = transtrn(input,'INF_',trimn(' '));

Of if you know the values are always at the beginning of the string then use SUBSTR.

desired_result = input ;

if desired_result =: 'INF_' then desired_result = substr(desired_result,5) ;