topic Re: Return word number in a string in New SAS User

Return word number in a string

DumDum — Fri, 09 May 2025 23:14:44 GMT

I am attempting to find the location of a word within a string, not the character position, but where in the string the word appears. For example, the following code:

data example;
    input string $50.;
    word = 'word';
    position = find(string, word);
    datalines;
This is a sample string containing the word.
Another example without the keyword.
The word appears here.
;
run;

Has the output of

However, the position I am seeking is:
OBS1 position 8
OBS2 position 0

OBS3 position 2

Reason being the actual data will have wildcards scattered within the string that could be replaced with any number of characters and this value will later be used in a scan function to find specific matches in a delimited variable value.

Hope that makes sense...

Re: Return word number in a string

quickbluefish — Sat, 10 May 2025 00:35:24 GMT

data example;
    input string $50.;
    word = 'word';
    position = findw(compress(string, '.,;!?'), word, ' ', 'e');
    datalines;
This is a sample string containing the word.
Another example without the keyword.
The word appears here.
;
run;

Note that this will only find whole words, so, for example, it won't find "word" in "keyword". Annoyingly, it also doesn't consider a word that ends with punctuation, like "word." to be a word either - the compress part is a workaround for that.

Re: Return word number in a string

Tom — Sat, 10 May 2025 00:44:57 GMT

Annoyingly, it also doesn't consider a word that ends with punctuation, like "word." to be a word either

That is what the delimiter parameter is for. Also consider the use of the I and T options. And the S option is useful with T to make sure that the space does to get removed from the delimiter list.

524  data example;
525    input string $50.;
526    word = 'word';
527    position = findw(string, word, '.?!,:;','site');
528    put position= string=;
529  datalines;

position=8 string=This is a sample string containing the word.
position=0 string=Another example without the keyword.
position=2 string=The word appears here.

Re: Return word number in a string

Ksharp — Sat, 10 May 2025 03:11:47 GMT

Using E option to find nth word ,not the position of word,

and using P option to add punctuation as a delimiter character.

data example;
    input string $50.;
    word = 'word';
    position = findw(string,strip(word),' ','pe');
    datalines;
This is a sample string containing the word.
Another example without the keyword.
The word appears here.
;
run;
proc print;run;