topic Re: compress few strings in a large string in SAS Programming

compress few strings in a large string

lokeshbm — Thu, 28 Jan 2016 05:52:23 GMT

Hi,

Is there a function or a loop which can compress multiple strings in one large string.

for ex: if (salary >= 0 & salary <= 25000) then salcat_1 = "C";

output should look like : salary >= 0 & salary <= 25000 salcat_1 = "C"

Appreciate your help on this.

Re: compress few strings in a large string

ad123123 — Thu, 28 Jan 2016 07:56:02 GMT

Hi,

Could you please explain criterian for removing multiple strings?

if (salary >= 0 & salary <= 25000) then salcat_1 = "C";

In the string above , you are eliminating "if" and "then". Will this be the same always? Every time we need to eliminate only "If" and "else"? Please be more explicit.

Re: compress few strings in a large string

lokeshbm — Thu, 28 Jan 2016 08:47:53 GMT

I am importing a file which has few strings {ex: IF THEN ELSE SELECT etc} ,I now wanted to remove those strings{not pre post text,only the above mentioned words} from the dataset .

As i have split those in to observations,I am in serach of a best approach/logic, Just posted a basic example below.

input : if (salary >= 0 & salary <= 25000) then salcat_1 = "C";

Output : salary >= 0 & salary <= 25000 salcat_1 = "C"

Re: compress few strings in a large string

RW9 — Thu, 28 Jan 2016 10:13:11 GMT

What about the ()? What about other characters? If you just want to remove those then tranwrd() or compress() can do it:

data want;
  set have;
  string=tranwrd(string,"if","");
  string=tranwrd(string,"then","");
run;

However I question what you are doing, you appear to be reading in code, which can make parsing more difficult than just removing some characters, e.g.

if (abc="I go on holiday if I booked it) then result="Sometimes (always)";

will result in:

(abc="I go on holiday I booked it) result="Sometimes (always)";

Personally, I would look at other methods to do this, including looking at the documentation - you do have that?

Re: compress few strings in a large string

Reeza — Thu, 28 Jan 2016 16:22:36 GMT

@lokeshbm wrote:

I now wanted to remove those strings{not pre post text,only the above mentioned words} from the dataset .

I can't see any list of mentioned words or characters. You need to define those.

Then you can use a combination of translate or transwrd functions.

Regular expressions can probably help as well.