topic Re: Encoding when reading data from api in SAS Programming

Encoding when reading data from api

rudfaden — Tue, 09 Jan 2018 12:29:55 GMT

How do I get the correct encoding for this data i am retrieving form with an api? The Danish letters æ,ø,å does not show corret

filename src temp encoding="utf-8";
proc http
     method="GET"
     url="%nrstr(https://dawa.aws.dk/adresser?q=Ærøgade&struktur=mini&per_side=1&format=csv)"
     out=src;
    run;
	data _null_;
	infile src;
	input;
	put _infile_;
	run;
	data test ;
	infile src encoding="utf-8" dlm=',' missover dsd firstobs=2 ;
	length id $ 36 status $1 vejkode $4 vejnavn $20 adresseringsvejnavn $20 husnr $10 etage $10 dor $10 supplerendebynavn $10 postnr $4 postnrnavn $20 kommunekode $4 adgangsadresseid $36 x 8 y 8 ;  
   input id $ status $ vejkode $ vejnavn $ adresseringsvejnavn $ husnr $ etage$ dor $ supplerendebynavn $ postnr $ postnrnavn $ kommunekode $ adgangsadresseid x y;
run;

Data example

id,status,vejkode,vejnavn,adresseringsvejnavn,husnr,etage,dÃ¸r,supplerendebynavn,postnr,postnrnavn
,kommunekode,adgangsadresseid,x,y
0a3f50b3-135b-32b8-e044-0003ba298018,1,0365,Ã†rÃ¸gade,Ã†rÃ¸gade,2,,,,5960,Marstal,0492,0a3f5088-8b
75-32b8-e044-0003ba298018,10.51420646,54.85857924

Re: Encoding when reading data from api

RW9 — Tue, 09 Jan 2018 13:13:57 GMT

What is your SAS session encoding? This is normally setup when SAS starts up, you would need to change it there.

Re: Encoding when reading data from api

rudfaden — Wed, 10 Jan 2018 09:14:43 GMT

That is not possible. I need a solution where the encoding is converted to my sas encoding,

Re: Encoding when reading data from api

RW9 — Wed, 10 Jan 2018 09:21:00 GMT

I don't believe that is possible. If your SAS session is Wlatin, and the data coming in has UTF-8 or 16, which both have far more characters in their sets, then "converting" data is not possible as the wlatin charset is too small to handle it. This is what you are seeing now, where the character set just says I don't know and puts a strange symbol.

Re: Encoding when reading data from api

rudfaden — Mon, 18 Mar 2019 10:59:01 GMT

I found a solution. Basically you save the file with utf-8 and then open it with wlatin.

filename resp "C:\Temp\map.json";
    filename minmap 'C:/Temp/gogle_distance.map';

    proc http
     url="https://maps.googleapis.com/maps/api/directions/json?origin=&home.%nrstr(&destination=)&work."
     method= "GET"
     out=resp;
    run;

    %** trick der gøre at JSON filen encodes med wlatin;
    filename resp "C:\Temp\map.json" encoding='utf-8';
    filename resp1 "C:\Temp\map1.json" encoding='wlatin1';

    data _null_;
      infile resp;
      file resp1;
      input;
      put _infile_;
    run;

Re: Encoding when reading data from api

rudfaden — Thu, 08 Feb 2018 08:31:09 GMT

I am not talking about converting the whole charset to from utf-8 to wlatin. Only those that are in wlatin