Count number of Regular Expression Matchees in SAS-CodePudding

I would appreciate any help in trying to count the number of matches of a regular expression function in SAS. I need to count the number of times a financial transaction is present in a text string.

I have been able to write the expression to find the transactions but have not been able to count how many times it may appear in a string. For illustration, I have created the following table:

DATA Transactions;
    infile datalines dlm='  ';
    INPUT ID STRING $40.;
    DATALINES;
1   One transaction R$250
2   One instance of R$ 250 and one of R$ 200
3   Nothing to see here R$
;
RUN;

And I have created a variable which can check if a transaction is present.

DATA Want;
    Set Transactions;
    QtdMatch = PRXMATCH("/ R\$ ?\d /",String);
RUN;

However, what I need to have is a variable counting the number os transactions. So, in the first string the variable would read 1, for the 2nd string 2 and for the final example 0. My strings could contain up to 50 transactions which would need to be counted.

CodePudding user response：

You can iterate through all the matches using call prxnext. This solution is adapted from the SAS documentation on CALL PRXNEXT.

data want;
    set transactions;

    regex        = prxparse("/ R\$ ?\d /");
    stop         = length(string);
    transactions = 0;

    call prxnext(regex, 1, stop, string, position, length);

    do while (position > 0);
        transactions 1;
        call prxnext(regex, 1, stop, string, position, length);
    end;
run;

Output:

ID  STRING                                      transactions
1   One transaction R$250                       1
2   One instance of R$ 250 and one of R$ 250    2
3   Nothing to see here R$                      0