How do i grab only the count, xx, out of this xx (xx.xx)?

[Hello_there]

data have1;
infile datalines dsd dlm=",";
*length value $200;
	input layer $ sort value $ 20.;
	value_n=input(substr(value, 1, 3), best.);
datalines;
001, 1, 20 (13.00)
002, 5, 35 (12.08)
003, 1, 267 (21.00)
004, 2, 28 (72.00)
005, 2, 86 (2.23)
006, 3, 24 (13.69)
007, 4, 21 (29.14)
008, 5, 7 (9.02)
009, 5, 56 (17.00)
010, 2, 21 (56.21)
;
run;

Hi,

Value is a character variable that is in the form xx (xx.xx). I'm trying to turn the xx into a new numerical variable called value_n so that i can rank it.

I tried using the substr function, but as you can see, there is one 1 digit number, and 1 3 digit number that causes the problem when i extend the parameters to the first 3. When this happens the row w/ a 7 gets a missing value.

Is there another way to do this?

[mkeintz · Accepted solution]

Don't use substr, because it retrieves a fixed count of characters, which creates the problem you have when you really need to accommodate a variable number of characters.  Use the SCAN function which allows you to retrieve the n'th "word" from a string - in your case you want the 1st word, where words are delimited by blanks.

I.e. change

	value_n=input(substr(value, 1, 3), best.);

to

    value_n=input(scan(value,1,' '),best.);

Change the second argument of SCAN to 2 to get the second "word", 3 for the 3rd, etc.

But if you always want the last word and you have a variable number of words in the character variable, use a negative offset, as in:

    value_last=input(scan(value,-1,' '),best.);

[Hello_there]

Thanks, worked like a charm!