topic Re: Searching text substring issue? in SAS Programming

Searching text substring issue?

ybz12003 — Tue, 05 Dec 2017 16:07:23 GMT

Hello,

I have a problem when searching the text substring "NOS". My codes is list below.

data have;

length label $ 120;

input label &;

cards;

5-AMINOSALICYLIC ACID

ANTIEMESIS-NOS

GI/PROKINETIC AGENTS

FLINTSTONES MULTIVITAMINS- NOS

FRUCTUS CYNOSBATI, HERB

INOSITOL

INOSITOL NIACINATE

LACTOBACILLUS RHAMNOSUS

MIDOL NOS

GI/ANTIEMETICS

NOS- ANTACID

;

run;

data want;

set have;

Labelpos1=find(label,"NOS","i");

run;

data nos;

set want;

where Labelpos1 ne 0;

run;

The result I am looking for is:

ANTIEMESIS-NOS

FLINTSTONES MULTIVITAMINS- NOS

MIDOL NOS

NOS- ANTACID

However I found some other names containing "NOS" in the middle of the text, I don't want to include them. The list is shown below. Please advice how to fix it. Thanks.

The list I don't want:

5-AMINOSALICYLIC ACID

FRUCTUS CYNOSBATI, HERB

INOSITOL

INOSITOL NIACINATE

LACTOBACILLUS RHAMNOSUS

Re: Searching text substring issue?

Shmuel — Tue, 05 Dec 2017 16:18:03 GMT

Use function findw instead of find:

data want;
set have;
     if findw(upcase(label), 'NOS' , ' -');
run;

Re: Searching text substring issue?

HB — Tue, 05 Dec 2017 16:33:03 GMT

It looks like you want NOS in three conditions:

1. Space before and after NOS

2. Hyphen before -NOS

3. Hyphen after NOS-

So one option (there are probably cleaner options) would be to modify your existing code as follows:

data want;
set have;
Labelpos1=find(label," NOS ","i");
Labelpos2=find(label,"-NOS","i");
Labelpos3=find(label,"NOS-","i");
run;

data nos;
set want;
where Labelpos1 ne 0 or Labelpos2 ne 0 or Labelpos3 ne 0;
run;

That yields

The SAS System

label	Labelpos1	Labelpos2	Labelpos3
ANTIEMESIS-NOS	0	11	0
FLINTSTONES MULTIVITAMINS- NOS	27	0	0
MIDOL NOS	6	0	0
NOS- ANTACID	0	0	1

which you can clean up.

EDIT: See? I told you there was cleaner. @Shmuel 's solution is MUCH cleaner. Use it.

Re: Searching text substring issue?

ybz12003 — Fri, 08 Dec 2017 14:35:28 GMT

Hi Shmuel:

Your code is very neat. However, it doesn't work in my actual data. Thanks for spending the time helping me.