topic Re: Spliting a string with all upper case in SAS Programming

Spliting a string with all upper case

Rahul_SAS — Thu, 14 Nov 2019 13:22:41 GMT

Hi Experts,

I have a variable i.e. Var1 and want to split it as shown in Want column below:

In Var1 all values do not have Spaces between words so need to insert a space inbetween.

data ana_pres.have;
input Var1 $30.;
datalines;
ABCReport
MainUsers
A-ABCTestReport
Overview(ABC)
ABC-I2.5.3
Report+DOS
ABC/DS Report
;
run;

HAVE	WANT
ABCReport	ABC Report
MainUsers	Main Users
A-ABCTestReport	A-ABC Test Report
Overview(ABC)	Overview (ABC)
ABC-I2.5.3	ABC -I2.5.3
Report+DOS	Report + DOS
ABC/DS Report	ABC / DS Report

Thanks

Re: Spliting a string with all upper case

gamotte — Thu, 14 Nov 2019 13:54:59 GMT

Hello,

The purpose of this forum is to give/get technical help regarding the SAS language.

Here your problem lacks clear specifications. You have to design unambiguous rules

to decide when and where spaces should be inserted and then, users of this forum

can help you implement your design in SAS. Your example does not make

that clear at all.

For instance, why "Overview" gives "Overview"and not "O verview" ?

Why "ABCReport" gives "ABC Report" and not "ABCR eport" ?

What should be done with non alphabetical characters surroundig uppercase

strings : "Overview(ABC)" gives "Overview( ABC )" or "Overview (ABC) " ?

Re: Spliting a string with all upper case

s_lassen — Thu, 14 Nov 2019 14:19:45 GMT

This probably calls for PRX:

data have;
input Var1 $30.;
datalines;
ABCReport
MainUsers
A-ABCTestReport
Overview(ABC)
ABC-I2.5.3
Report+DOS
ABC/DS Report
;
run;

data want;
  set have;
  prxid1=prxparse('s/([A-Z,a-z])([A-Z][a-z])/$1 $2/');
  prxid2=prxparse('s/([+,\/])/ $1 /');
  prxid3=prxparse('s/\(/ (/');
  call prxchange(prxid1,-1,Var1);
  call prxchange(prxid2,-1,Var1);
  call prxchange(prxid3,-1,Var1);
run;

The first regular expression inserts a space when you have a letter followed by a capital letter, followed by a non-capital letter.

The second inserts spaces around "+" and "/" (the latter is escaped with a backslash in the search string).

The third inserts a blank before "(".

But I am not sure what to do about the "-", in one case you precede "-" with a blank, in the other case you do not.

Re: Spliting a string with all upper case

yabwon — Thu, 14 Nov 2019 14:27:37 GMT

Hi,

Try this one (not perfect, but 90%)

data have;
  input Var1 $30.;
datalines;
ABCReport
MainUsers
A-ABCTestReport
Overview(ABC)
ABC-I2.5.3
Report+DOS
ABC/DSReport
;
run;

data want;
  set have;
  length Var2 $ 60;
  keep Var1 Var2;

  array letters[60] $ 1 _temporary_;
  array markers[3] $ 1 _temporary_ ('=' '+' '/');

  i = 0;
  do until(letters[i] = " ");
    i+1;
    letters[i] = char(Var1,i);
  end;
  i+(-1);

  do i = 1 to i;
    select;
      when ( ('A' <= letters[i] <= 'Z') and ('A' <= letters[i+1] <= 'Z') )
        do; 
          Var2 = strip(Var2) || letters[i] ;
        end;
      when ( ('a' <= letters[i] <= 'z') and not('a' <= letters[i+1] <= 'z') )
        do; 
          Var2 = strip(Var2) || letters[i] || " " || letters[i+1];
          i + 1;
        end;
      when ( ('A' <= letters[i] <= 'Z') and ('a' <= letters[i+1] <= 'z') )
        do; 
          Var2 = strip(Var2) || " " || letters[i] ||  letters[i+1];
          i + 1;
        end;
      otherwise Var2 = strip(Var2) || letters[i] ; 
    end;
  end;

  do i = 1 to dim(markers);
    Var2 = compbl(tranwrd(Var2, markers[i], " "||markers[i]||' '));
  end;
run;

But since collapsing spaces is irreversible function having general solution may not be doable 🙂

All the best

Bart

Re: Spliting a string with all upper case

Ksharp — Fri, 15 Nov 2019 13:35:03 GMT

This could give you a start.

data have;
input Var1 $30.;
datalines;
ABCReport
MainUsers
A-ABCTestReport
Overview(ABC)
ABC-I2.5.3
Report+DOS
ABC/DS Report
;
run;

data want;
 set have;
 want=left(compbl(prxchange('s/([A-Z][a-z]+)/ \1 /',-1,var1)));
 want=left(compbl(prxchange('s/(\W)/ \1 /',-1,want)));
run;