topic Re: Find and replace a substring in SAS Programming

Find and replace a substring

Stefy67 — Wed, 04 Oct 2017 14:06:45 GMT

I have this problem: some values in VAR1 (see example) have special characters ( À Á Â Ã Ä ). I want to find these special characters in a string and replace it with the correct letter (A) as showed in VAR2 of the example.

Var1	Var2
SIAHMED	SIAHMED
DANDREÁ	DANDREA
PELLIÇANO	PELLICANO
MATÃ	MATA
DONZELLA	DONZELLA
DAGOSTINO	DAGOSTINO
ÔTTINA	OTTINA

I tried this instruction

var2 = TRANWRD (var1, "Á ", "A"); and it works but if I include more than one target it doesn't

var2 = TRANWRD (var1, "Á" or "Ã", "A");

Waiting for your suggestion

Thanks

Re: Find and replace a substring

Astounding — Wed, 04 Oct 2017 14:31:36 GMT

You would do better to switch from TRANWRD to TRANSLATE:

http://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=9.4_3.2&docsetId=lefunctionsref&docsetTarget=p05ww22zp7lcg3n1bjk7v93tscyo.htm&locale=en

That does exactly what you are asking for.

Re: Find and replace a substring

mkeintz — Wed, 04 Oct 2017 14:44:24 GMT

Check out the TRANSLATE function, which expression like the below, which converts all vowels to X.

name=translate(name,'XXXXX','AEIOU');

So presumably you could use

var2=translate(var1,'AAAAA','ÀÁÂÃÄ');

Just be sure that the second argument has as many iterations of the common value ('A') as there are characters in the 3rd argument.

Re: Find and replace a substring

Ksharp — Thu, 05 Oct 2017 07:16:56 GMT

Calling @Patrick

Re: Find and replace a substring

Stefy67 — Thu, 05 Oct 2017 09:20:35 GMT

Referring to my previous question, I write two lines together as below but the instruction doesn’t work.
Do I have to split for different dataset one instruction at a time?

Thank you
Stefy67

data tre;
set due;
nomenew3 = translate(nomenew2,'AAAAAAAAAAA','ÀÁÂÃÄÅÆĀĂĄǍ');
cognomenew3 = translate(cognomenew2,'AAAAAAAAAAA','ÀÁÂÃÄÅÆĀĂĄǍ');

run;

Re: Find and replace a substring

Patrick — Thu, 05 Oct 2017 14:21:50 GMT

@Stefy67

Below one way to go. The approach works only for single byte character sets and requires one of the more recent SAS versions.

data due;
  nomenew2 = 'À à ÁÂÃÄÅ ÆAAAA è';
  cognomenew2 = 'ÔTTINA ÀÁÂÃÄÅÆAAAA';
run;

proc format;
   invalue $norm_to_A (default=256) 's/[ÀÁÂÃÄÅÆäà]/A/'   (regexpe) = _same_ other=_same_;
   invalue $norm_to_O (default=256) 's/[Ô]/O/'             (regexpe) = _same_ other=_same_;
   invalue $norm_to_E (default=256) 's/[è]/E/'             (regexpe) = _same_ other=_same_;
run;

data tre;
  set due;
  array c_vars {*} nomenew2 cognomenew2;
  do _i=1 to dim(c_vars);
    do _infmt='$norm_to_A','$norm_to_O','$norm_to_E';
  	 c_vars[_i]=inputc(c_vars[_i],_infmt);
    end;
  end;
run;

Re: Find and replace a substring

Stefy67 — Thu, 05 Oct 2017 14:11:35 GMT

Thanks Patrick

I solved the problem with your suggestions.