topic Re: index in SAS Programming

index

HeatherNewton — Mon, 07 Mar 2022 10:42:36 GMT

options nodate nostimer ls=78 ps=60;
data _null_;
   length a b $14;
   a='ABC.DEF (X=Y)';
   b='X=Y';
   q=index(a,b);
   w=index(a,trim(b));
   put q= w=;
run;

I dont understand the answer, why is q=0 and w=10?

Re: index

Kurt_Bremser — Mon, 07 Mar 2022 10:54:18 GMT

Always keep in mind that character variables are padded to their defined length with blanks, and in the first call of INDEX, those blanks are there in variable b and would need to also be present in a (which they are not, of course).

In the second call to INDEX, the TRIM function removes the trailing blanks, and the remaining string is found.

Re: index

PeterClemmensen — Mon, 07 Mar 2022 10:56:31 GMT

This is Example 2 from the Index Function Documentation.

data _null_;
   length a b $14;
   a='ABC.DEF (X=Y)';
   b='X=Y';
   q=index(a, b);
   w=index(a, trim(b));
   put q= w=;
run;

What you want to notice here is the Length Statement, giving both a and b a length of $14. Meaning that SAS pads both a and b with trailing blanks. The code below gives the same as above without the Length Statement.

data _null_;
   a='ABC.DEF (X=Y) ';
   b='X=Y           ';
   q=index(a, b);
   w=index(a, trim(b));
   put q= w=;
run;

This means that in index(a, b) looks for the string 'X=Y ' (including trailing blanks) in the string 'ABC.DEF (X=Y) '. This returns zero because it's not there. Index(a, trim(b)) looks for 'X=Y' (without trailing blanks) in the same string. This returns 10.

Re: index

PaigeMiller — Mon, 07 Mar 2022 10:58:11 GMT

Variable B has length 14. It has the value X=Y plus 11 blanks. This string of X=Y followed by 11 blanks does not appear in A.

When you do the TRIM then the value is X=Y without any trailing blanks, which does appear in A.

To fix this use

length a $14;

or even remove the LENGTH statement entirely as it is not needed in this data set.

Re: index

Patrick — Mon, 07 Mar 2022 11:07:00 GMT

The secret is hidden in the length statement.

Because this length statement defines b with a length of $14 the index() function will use the fully expanded value as below (blanks are also just characters):

Because this is not a match the index() function returns a value of zero.

If you use the trim function then the index() function will see the following (trailing blanks removed):

Now it's a match and that's why the index() function returns a value of 10.

Does that explain it to you?

Re: index

HeatherNewton — Tue, 08 Mar 2022 02:48:14 GMT

Yes thank you all for the clear explanation!

Re: index

SASKiwi — Tue, 08 Mar 2022 03:18:44 GMT

@HeatherNewton - If the answers fix your problem, it's good practice to update the post as answered 🙂.