topic Re: Multiple CSVs in FILENAME ZIP via INFILE FILEVAR? in SAS Programming

Multiple CSVs in FILENAME ZIP via INFILE FILEVAR?

Junyong — Fri, 19 Mar 2021 05:41:32 GMT

FILEVAR allows to INFILE multiple CSVs with the same structure. How can I use this if the CSVs are in FILENAME ZIP? For example,

filename i "wid_all_data.zip";

proc http url="https://wid.world/bulk_download/wid_all_data.zip" out=i;
run;

filename i zip "wid_all_data.zip";

data i;
	i=dopen("i");
	do j=1 to dnum(i);
		k=cats('i("',dread(i,j),'")');
		output;
	end;
run;

data j;
	set i;
	infile l filevar=k dlm=";" firstobs=2;
	input (country variable percentile)(:$16.) year value age pop :$16.;
run;

SAS does not use k=i("WID_data_AN.csv") as FILEVAR here.

Re: Multiple CSVs in FILENAME ZIP via INFILE FILEVAR?

qoit — Fri, 19 Mar 2021 06:03:09 GMT

Try the below great article, apologies as I am unbale to access URL links via my SAS client but try the below link:

https://blogs.sas.com/content/sasdummy/2015/05/11/using-filename-zip-to-unzip-and-read-data-files-in-sas/

Just add the below line:

filename inzip ZIP "c:\projects\data.zip";
 
/* Read the "members" (files) from the ZIP file */
data contents(keep=memname isFolder);
 length memname $200 isFolder 8;
 fid=dopen("inzip");
 if fid=0 then
  stop;
 memcount=dnum(fid);
 do i=1 to memcount;
  memname=dread(fid,i);
  /* check for trailing / in folder name */
  isFolder = (first(reverse(trim(memname)))='/');
  output;
 end;
if upcase(scan(memname,-1,".")) = CSV; /* Flag all CSV files */
 rc=dclose(fid);
run;

Re: Multiple CSVs in FILENAME ZIP via INFILE FILEVAR?

Junyong — Fri, 19 Mar 2021 06:23:53 GMT

Can't I INFILE multiple CSVs in a ZIP without unzipping them?

Re: Multiple CSVs in FILENAME ZIP via INFILE FILEVAR?

Kurt_Bremser — Fri, 19 Mar 2021 08:05:28 GMT

Sadly, this is not possible. To read a file in a ZIP archive, you need to create a file reference to the ZIP, and use FILEREF(MEMBER) syntax as the file specification in the INFILE statement. The variable used in the FILEVAR= option, OTOH, must hold a physical filename.

You will have to import each file separately, and concatenate the results.