Extract symbol or number from text

[Johnlee]

Dear all,

How to pick up a symbol or/and number in a paragraph? for example, how to extract "#" or "678" or "#  678" from the following data named as text? The attached dataset can be read into SAS in the following code.

DATA text;

INFILE "C:\text.txt"

dlm="," lrecl=5000;

informat var1 $256.; 

format var1 $256.;   

input var1;

RUN;

thanks!

John

[Reeza]

Look into the compress function for starters. Regular expressions will also get you there, but a bit harder to understand off the start.

data want;

set text;

var2=compress(var1, , 'a');

run;

proc print data=want;

run;

[Ksharp · Accepted solution]

COMPRESS() is enough powerful for this .

data want;

set text;

var2=compress(var1,'#' , 'kd');

run;

Xia Keshan

[Johnlee]

Many thanks Reeza and Keshan!!