topic Re: how to read file from unix directory in SAS Programming

how to read file from unix directory

Aexor — Tue, 04 May 2021 13:20:10 GMT

Hi All,

Help needed in below question:

A unix directory contains 30 csv files . i want to read the 27th file csv.

Please help.

Re: how to read file from unix directory

Kurt_Bremser — Tue, 04 May 2021 14:40:06 GMT

In a data step, open the directory (assign a file reference with the FILENAME function, then use that in a DOPEN call).

Now you can loop over the contents of the directory (DREAD, up to the value returned by the DNUM function), and assign the returned filename to a macro variable (CALL SYMPUTX) when the counter hits 27.

Re: how to read file from unix directory

Tom — Tue, 04 May 2021 14:58:25 GMT

@Aexor wrote:

Hi All,

Help needed in below question:

A unix directory contains 30 csv files . i want to read the 27th file csv.

Please help.

You need to provide a lot more detail about what you are doing.

Do you just want to find the name of the 27th file? How are you ordering the files to determine which is the 27th?

Do you want to actually read the file once you have found its name? Do you know the fields that are supposed to be in the CSV file? Or will you have to use a tool such as PROC IMPORT to guess how to read the file into data?

Re: how to read file from unix directory

Aexor — Tue, 04 May 2021 15:09:33 GMT

Hi Tom,

Thanks for checking.
Yes the files are in sequential order a.txt, b.txt and so on .so I want to read the 27th file and filelds are same in all files e.g name,class,number

Re: how to read file from unix directory

Kurt_Bremser — Tue, 04 May 2021 15:25:05 GMT

You will have to read all filenames into a dataset, and sort that by name, then you can get the 27th with a POINT= option:

data fnames;
length dref $8 fname $200;
rc = filename(dref,"/folders/myfolders");
did = dopen(dref);
if did
then do i = 1 to min(dnum(did),27);
  fname = dread(did,i);
  output;
end;
if i le 27 then putlog "Not enough files!";
keep fname;
run;

proc sort data=fnames;
by fname;
run;

data _null_;
pt = 27;
set fnames point=pt;
call symputx('myfile',fname);
stop;
run;

%put &myfile.;