DATA Step, Macro, Functions and more

Extracting words from string

Reply
Contributor
Posts: 50

Extracting words from string

Hello All

What is the effective way to extract two words from the following string!? ( I need to extract Day 1)

I can use substr(string,5,5), but sometimes it can be Day 100 or W1 Day 25 (Study 2) in which i need to extract Day 25

I can also use scan function and extract 2 and  3 words and concatenate ! Is there any function which is more robustic?

W10 Day 1 (Study 1)

W1 Day 100 (Study x)

W Day 25

Thanks for all your insights!

J

Super Contributor
Super Contributor
Posts: 3,174

Re: Extracting words from string

Posted in reply to SASSLICK001

SCAN works well, either supplying a positive-value argument (from start-location) or negative-value (from end-location) -- will depend on the result/requirement, however there are likely multiple SAS functions that can accomplish the task.  Similarly, various SAS concatenation functions as well will apply, such as CATX, or using TRIM, possibly even COMPRESS, depending on the rqmt.

Scott Barry
SBBWorks, Inc.

Super User
Posts: 10,035

Re: Extracting words from string

Posted in reply to SASSLICK001
data have;
input x $40.;
cards;
W10 Day 1 (Study 1)
W1 Day 100 (Study x)
W Day 25
;
run;
data want;
 set have;
if prxmatch('/\s+Day\s+\d+\s+/i',x) then 
want=prxchange('s/.*\s+(Day\s+\d+)\s+.*/$1/i',-1,x);
run;

Xia Keshan

Super User
Posts: 10,035

Re: Extracting words from string

Posted in reply to SASSLICK001

data want;

set have;

retain pid;

if _n_ eq 1 then pid=prxparse('/\s+Day\s+\d+\s+/i');

call prxsubstr(pid, x, position, length);

if position ne 0 then want = substr(x, position, length);

drop pid position length;

run;

Contributor
Posts: 50

Re: Extracting words from string

Thanks Keshan for your reply,

whan about if we have W1 Screening

Super User
Posts: 10,035

Re: Extracting words from string

Posted in reply to SASSLICK001

It will return a missing value .

Ask a Question
Discussion stats
  • 5 replies
  • 624 views
  • 0 likes
  • 3 in conversation