topic Re: compress or left trim in SAS Procedures

compress or left trim

archibald — Mon, 08 Feb 2016 02:08:26 GMT

Hi all,

I have a character variable ID is as follows :

1002-004567-UC

1002-000062-UC

1002-239874-UC

I would like to transform the ID variable such as:

4567

239874

What would be the appropriate syntax to use?

Thank you.

Re: compress or left trim

stat_sas — Mon, 08 Feb 2016 02:33:18 GMT

data have;
input ID $20.;
datalines;
1002-004567-UC
1002-000062-UC
1002-239874-UC
;

data want(drop=id rename=id1=id);
set have;
ID1=input(scan(ID,2),8.);
run;

Re: compress or left trim

Reeza — Mon, 08 Feb 2016 03:06:34 GMT

Do you want the variable to be character or numeric?

The solution above generates a numeric ID. In general ID's should be character to avoid accidental mathematical issues and precision issues in merging.

SCAN() isolates the middle term

INPUT() converts to a number, so it removes the leading zero's

PUT() converts it back to a character, -l, left aligns the variable.

ID_VAR = put(input(scan(var, 2, "-"), 8.), 8. -l);

Re: compress or left trim

PGStats — Mon, 08 Feb 2016 03:58:03 GMT

You could also use pattern matching

data ids;
input idStr :$16.;
datalines;
1002-004567-UC
1002-000062-UC
1002-239874-UC
;

data want;
set ids;
id = prxchange("s/.*-0*([1-9]\d*)-.*/\1/o", 1, idStr);
run;

proc print data=want noobs; run;

The pattern reads: Find any string, followed by a dash, followed by any number of zeros, ( followed by a digit between 1 and 9, followed by any number of digits ), followed by a dash, followed by anything. Keep the string part matched within the parentheses.

Re: compress or left trim

archibald — Mon, 08 Feb 2016 22:56:37 GMT

PGStats,
This works great! Thanks!!
Also say the ID variable was as follows:
1002-004567UC
1002-000062UC
1002-239874UC
which part of the code should I modify to get the same result
ID
4567
62
239874

Re: compress or left trim

archibald — Mon, 08 Feb 2016 23:01:22 GMT

sat_sas,
the code provided does not work for me. I got the following error message :Invalid argument to function INPUT

Re: compress or left trim

archibald — Mon, 08 Feb 2016 23:06:15 GMT

Thanks Reeza for your reply.
I am not sure this is on my end, but I was not successful using this code. I got an error message Invalid argument to function INPUT
perhaps I am missing something...

Re: compress or left trim

PGStats — Mon, 08 Feb 2016 23:08:34 GMT

Replace the pattern with "s/.*-0*([1-9]\d*)\D.*/\1/o"

It would actually work also with your original question. It says that the number field ends when a non number character is found.

Re: compress or left trim

PGStats — Mon, 08 Feb 2016 23:11:51 GMT

I guess you could just as well use "s/.*-0*([1-9]\d*).*/\1/o" which would work even if the number goes to the end of the string.

Re: compress or left trim

archibald — Tue, 09 Feb 2016 05:07:45 GMT

great! thanks again!