topic Re: How do I import a .json file with headers designation in SAS Programming

How do I import a .json file with headers designation

edesantis — Mon, 11 Jul 2016 15:37:37 GMT

I have a .json file I need to read into sas. This file has a row with headers and then has the rows of data listed below. Can anyone assist in what the sas code to import this data correctly would look like?

{"headers":
[["Grad Rates Over Time","4-Year Rate 2000","4-Year Rate 2001","4-Year Rate 2002","4-Year Rate 2003","4-Year Rate 2004","4-Year Rate 2005","4-Year Rate 2006","4-Year Rate 2007","4-Year Rate 2008","4-Year Rate 2009","4-Year Rate 2010"]],
"rows":
[["California State University-Bakersfield","11.5%","13.3%","14.1%","14.1%","14.6%","12.9%","17.2%","14.9%","14.1%","15.0%","16.0%"],
["California State University-East Bay","14.5%","16.7%","16.3%","10.8%","16.0%","16.5%","14.5%","16.0%","14.9%","10.8%","10.4%"],
["California State University-San Bernardino","11.6%","10.0%","12.4%","10.0%","12.2%","13.3%","13.1%","12.9%","11.3%","9.4%","12.6%"],["California State University-Stanislaus","18.9%","21.5%","20.0%","18.6%","21.1%","21.6%","23.4%","20.4%","17.5%","18.2%","14.8%"],["Columbus State University","9.4%","8.1%","11.6%","11.2%","10.0%","11.2%","12.2%","12.3%","13.4%","12.1%","11.9%"],["CUNY John Jay College of Criminal Justice","10.8%","13.3%","17.1%","19.4%","20.7%","23.1%","21.2%","19.0%","21.0%","25.3%","22.3%"],["CUNY York College","5.7%","7.3%","6.3%","7.5%","6.0%","5.3%","3.7%","3.6%","4.0%","5.2%","7.8%"],["Francis Marion University","29.0%","18.2%","15.8%","20.0%","19.3%","20.0%","18.7%","19.8%","17.9%","16.4%","15.4%"],["Kean University","13.7%","16.1%","15.7%","17.3%","16.5%","18.8%","16.7%","20.7%","19.2%","18.9%","18.4%"],["Lamar University","16.8%","16.1%","17.9%","19.5%","18.1%","9.5%","11.1%","10.2%","10.9%","13.1%","11.0%"],["Northeastern Illinois University","1.6%","2.3%","2.8%","2.2%","2.8%","3.4%","3.8%","5.3%","3.9%","4.1%","5.0%"],["Texas A & M University-Corpus Christi","19.5%","18.4%","19.9%","16.4%","18.0%","18.6%","17.0%","20.6%","20.2%","24.3%","19.5%"],["Texas A & M University-Kingsville","8.4%","7.2%","9.6%","8.0%","14.0%","7.2%","10.8%","13.7%","13.0%","17.7%","16.8%"],["Texas Woman's University","21.1%","12.8%","21.2%","22.5%","27.6%","22.4%","23.1%","17.7%","22.6%","22.3%","23.3%"],["University of North Carolina at Pembroke","21.3%","18.7%","20.1%","14.5%","14.2%","16.4%","15.2%","14.8%","15.9%","15.4%","14.9%"],["Winston-Salem State University","21.8%","21.6%","19.4%","20.2%","18.1%","14.0%","13.1%","15.9%","14.8%","17.8%","17.3%"]]}

Re: How do I import a .json file with headers designation

Ksharp — Tue, 12 Jul 2016 02:25:58 GMT

There is a JSON Package in PROC DS2 . But for me not familiar with it , I would like to use data step:

data have;
infile '/folders/myfolders/have.json' recfm=n dsd dlm='[]{},:';
input temp : $100. @@;
if missing(temp) then group+1;
if findw(temp,'headers') or 
   find(temp,'"rows"') or 
   prxmatch('/^\s/',temp) then delete;
run;
proc transpose data=have out=temp(drop=_name_ group rename=(col1=label));
by group;
var temp;
run;
proc transpose data=temp out=want(drop=_name_);
var col:;
id label;
idlabel label;
run;
proc print noobs label;run;

Re: How do I import a .json file with headers designation

ChrisNZ — Tue, 12 Jul 2016 03:09:53 GMT

Like this?


data _null_;
  file "%sysfunc(pathname(WORK))\t.txt" lrecl=32000;
  infile cards pad truncover;
  input X $32000.;
  L=length(X);
  put X $varying32000. L; 
cards;
{"headers":
[["Grad Rates Over Time","4-Year Rate 2000","4-Year Rate 2001","4-Year Rate 2002","4-Year Rate 2003","4-Year Rate 2004","4-Year Rate 2005","4-Year Rate 2006","4-Year Rate 2007","4-Year Rate 2008","4-Year Rate 2009","4-Year Rate 2010"]],
"rows":
[["California State University-Bakersfield","11.5%","13.3%","14.1%","14.1%","14.6%","12.9%","17.2%","14.9%","14.1%","15.0%","16.0%"],
["California State University-East Bay","14.5%","16.7%","16.3%","10.8%","16.0%","16.5%","14.5%","16.0%","14.9%","10.8%","10.4%"],
["California State University-San Bernardino","11.6%","10.0%","12.4%","10.0%","12.2%","13.3%","13.1%","12.9%","11.3%","9.4%","12.6%"],["California State University-Stanislaus","18.9%","21.5%","20.0%","18.6%","21.1%","21.6%","23.4%","20.4%","17.5%","18.2%","14.8%"],["Columbus State University","9.4%","8.1%","11.6%","11.2%","10.0%","11.2%","12.2%","12.3%","13.4%","12.1%","11.9%"],["CUNY John Jay College of Criminal Justice","10.8%","13.3%","17.1%","19.4%","20.7%","23.1%","21.2%","19.0%","21.0%","25.3%","22.3%"],["CUNY York College","5.7%","7.3%","6.3%","7.5%","6.0%","5.3%","3.7%","3.6%","4.0%","5.2%","7.8%"],["Francis Marion University","29.0%","18.2%","15.8%","20.0%","19.3%","20.0%","18.7%","19.8%","17.9%","16.4%","15.4%"],["Kean University","13.7%","16.1%","15.7%","17.3%","16.5%","18.8%","16.7%","20.7%","19.2%","18.9%","18.4%"],["Lamar University","16.8%","16.1%","17.9%","19.5%","18.1%","9.5%","11.1%","10.2%","10.9%","13.1%","11.0%"],["Northeastern Illinois University","1.6%","2.3%","2.8%","2.2%","2.8%","3.4%","3.8%","5.3%","3.9%","4.1%","5.0%"],["Texas A & M University-Corpus Christi","19.5%","18.4%","19.9%","16.4%","18.0%","18.6%","17.0%","20.6%","20.2%","24.3%","19.5%"],["Texas A & M University-Kingsville","8.4%","7.2%","9.6%","8.0%","14.0%","7.2%","10.8%","13.7%","13.0%","17.7%","16.8%"],["Texas Woman's University","21.1%","12.8%","21.2%","22.5%","27.6%","22.4%","23.1%","17.7%","22.6%","22.3%","23.3%"],["University of North Carolina at Pembroke","21.3%","18.7%","20.1%","14.5%","14.2%","16.4%","15.2%","14.8%","15.9%","15.4%","14.9%"],["Winston-Salem State University","21.8%","21.6%","19.4%","20.2%","18.1%","14.0%","13.1%","15.9%","14.8%","17.8%","17.3%"]]}
run;

data _null_;
  infile "%sysfunc(pathname(WORK))\t.txt" lrecl=32000 firstobs=2 pad truncover;
  input X $32000.;
  X=tranwrd(X,'",','"n ');
  X=tranwrd(X,'"]','"n ');
  X=compress(X,'[],');  
  call symputx('names',X);
  stop;
run;

data WANT;
  infile "%sysfunc(pathname(WORK))\t.txt" lrecl=32000 firstobs=4 dlm='"},' pad truncover termstr=']';
  length &names. $64;
  input @'[' &names. ;
  putlog _N_= (&names)(=); 
  if lengthn(cats(of &names.));
run;

proc print noobs;
run;

Grad Rates Over Time	4-Year Rate 2000	4-Year Rate 2001	4-Year Rate 2002	4-Year Rate 2003	4-Year Rate 2004	4-Year Rate 2005	4-Year Rate 2006	4-Year Rate 2007	4-Year Rate 2008	4-Year Rate 2009	4-Year Rate 2010
California State University-East Bay	14.5%	16.7%	16.3%	10.8%	16.0%	16.5%	14.5%	16.0%	14.9%	10.8%	10.4%
California State University-San Bernardino	11.6%	10.0%	12.4%	10.0%	12.2%	13.3%	13.1%	12.9%	11.3%	9.4%	12.6%
California State University-Stanislaus	18.9%	21.5%	20.0%	18.6%	21.1%	21.6%	23.4%	20.4%	17.5%	18.2%	14.8%
Columbus State University	9.4%	8.1%	11.6%	11.2%	10.0%	11.2%	12.2%	12.3%	13.4%	12.1%	11.9%
CUNY John Jay College of Criminal Justice	10.8%	13.3%	17.1%	19.4%	20.7%	23.1%	21.2%	19.0%	21.0%	25.3%	22.3%
CUNY York College	5.7%	7.3%	6.3%	7.5%	6.0%	5.3%	3.7%	3.6%	4.0%	5.2%	7.8%
Francis Marion University	29.0%	18.2%	15.8%	20.0%	19.3%	20.0%	18.7%	19.8%	17.9%	16.4%	15.4%
Kean University	13.7%	16.1%	15.7%	17.3%	16.5%	18.8%	16.7%	20.7%	19.2%	18.9%	18.4%
Lamar University	16.8%	16.1%	17.9%	19.5%	18.1%	9.5%	11.1%	10.2%	10.9%	13.1%	11.0%
Northeastern Illinois University	1.6%	2.3%	2.8%	2.2%	2.8%	3.4%	3.8%	5.3%	3.9%	4.1%	5.0%
Texas A & M University-Corpus Christi	19.5%	18.4%	19.9%	16.4%	18.0%	18.6%	17.0%	20.6%	20.2%	24.3%	19.5%
Texas A & M University-Kingsville	8.4%	7.2%	9.6%	8.0%	14.0%	7.2%	10.8%	13.7%	13.0%	17.7%	16.8%
Texas Woman's University	21.1%	12.8%	21.2%	22.5%	27.6%	22.4%	23.1%	17.7%	22.6%	22.3%	23.3%
University of North Carolina at Pembroke	21.3%	18.7%	20.1%	14.5%	14.2%	16.4%	15.2%	14.8%	15.9%	15.4%	14.9%
Winston-Salem State University	21.8%	21.6%	19.4%	20.2%	18.1%	14.0%	13.1%	15.9%	14.8%	17.8%	17.3%

Re: How do I import a .json file with headers designation

PGStats — Tue, 12 Jul 2016 03:29:45 GMT

Proc JSON is only for writing json files. For importing a small (<32k) json file you could transform it into a CSV file first, like this:

/* Temporary file to hold CSV copy of data */
filename dum temp;

/* Extract headers and data rows from json file */
data _null_;
length str $32767 headers row $1000;
if not prx1 then prx1 + prxparse('/"headers":\s*\[\[(.*?)\]\]/i');
if not prx2 then prx2 + prxparse('/"rows":\s*\[(\[.*?\])\]/i');
if not prx3 then prx3 + prxparse('/\[.*?\]/');

/* Load entire file into a single string */
infile "&sasforum.\Datasets\JSON example.txt" end=done;
do until(done);
    input;
    str = catx(" ", str, _infile_);
    end;

/* Get and copy headers */
file dum;
if prxMatch(prx1, str) then do;
    headers = prxPosn(prx1, 1, str);
    put headers;
end;
  
/* Get and copy each row */
if prxMatch(prx2, str) then do;
    call prxPosn(prx2, 1, pos, len);
    start = pos;
    stop = pos + len - 1;
    call prxNext(prx3, start, stop, str, pos, len);
    do while(pos > 0);
        row = substr(str, pos+1, len-2);
        row = prxChange('s/"([0-9.%-]+?)"/\1/', -1, row);
        put row;
        call prxNext(prx3, start, stop, str, pos, len);
        end;
    end;
stop;
run;

/* Import temp file as a CSV file with headers */
proc import datafile="dum" out=test dbms=csv replace; 
guessingrows=max;
run;

Note that proc import doesn't like the percent character in numeric fields. Numeric fields without percent signs would be imported properly as numeric variables.

Re: How do I import a .json file with headers designation

ChrisNZ — Tue, 12 Jul 2016 05:17:21 GMT

@Ksharp Your variable names are truncated to 32.

Re: How do I import a .json file with headers designation

Ksharp — Tue, 12 Jul 2016 05:26:18 GMT

Chris,

That doesn't matter as long as you got that label and make sure variable name is unique .

Re: How do I import a .json file with headers designation

Ksharp — Tue, 12 Jul 2016 05:30:11 GMT

Chris,

If you want make variable name looks better , just change the ID variable.

data have;
infile '/folders/myfolders/have.json' recfm=n dsd dlm='[]{},:';
input temp : $100. @@;
if missing(temp) then group+1;
if  prxmatch('/^\s|^headers/',temp) then delete;
run;
proc transpose data=have out=temp(drop=_name_ rename=(col1=label));
by group;
var temp;
run;
proc transpose data=temp out=want(drop=_name_);
var col:;
id group;
idlabel label;
run;
proc print noobs label;run;

Re: How do I import a .json file with headers designation

ChrisNZ — Tue, 12 Jul 2016 05:46:31 GMT

I would argue it matters to have the name right rather than the label in the vast majority of cases, but to each their own. 🙂