Parse file name with SAS - sas

Question

I have a directory in which every week there is a new file created. names are like below:
file_w1.csv
file_w2.csv
file_w3.csv
What I need to do is pick up the latest file (based on modified date), then parse the 2 characters just before the file extension.
So in this case, I want 'w3' because I want to use this to know which week I am reporting for.
How can I do this in SAS?

Answer 1

An operating system independent technique would use SAS External File functions such as dopen, fopen and finfo to obtain information about a folder and it's items.
Consider this sample code that does a 'full dump' of available information whilst parsing C:\Temp on a Windows machine:
data _null_;
length dfileref fileref $8 folder $200;
rc = filename (dfileref, 'C:\Temp');
did = dopen(dfileref);
if did then do;
do index = 1 to doptnum(did);
featurename = doptname(did,index);
featurevalue = dinfo(did,featurename);
put index= featurename= featurevalue=;
if featurename = 'Directory' then folder = featurevalue;
end;
do dindex = 1 to dnum(did);
entryname = dread(did,dindex);
put dindex= entryname=;
rc = filename(fileref, cats(folder, '/', entryname));
fid = fopen (fileref); * if entry is another folder fid will be 0;
if fid then do;
do findex = 1 to foptnum(fid);
featurename = foptname(fid, findex);
featurevalue = finfo(fid, featurename);
put +2 findex= featurename= featurevalue=;
end;
fid = fclose(fid);
end;
rc = filename(fileref);
end;
did = dclose(did);
end;
rc = filename (dfileref);
run;
After examining the log you can pare down the code needed to gather specific desired information into a data set. You can then use SQL queries to further act upon the data:
data csv_files(keep=fullname lastmod where=(fullname like '%.csv'));
length dfileref fileref $8 folder $200;
folder = 'C:\Temp';
rc = filename (dfileref, folder);
did = dopen(dfileref);
if did then do;
do dindex = 1 to dnum(did);
entryname = dread(did,dindex);
rc = filename(fileref, cats(folder, '/', entryname));
fid = fopen (fileref);
if fid then do;
fullname = finfo(fid,'Filename');
lastmod = input(finfo(fid,'Last Modified'), datetime18.); format lastmod datetime18.;
output;
fid = fclose(fid);
end;
rc = filename(fileref);
end;
did = dclose(did);
end;
rc = filename (dfileref);
run;
proc sql;
create table csv_newest as
select *, scan(scan(fullname,-1,'_'),1,'.') as tag
from csv_files
where prxmatch ('/_.+\.csv$/', fullname)
having lastmod = max(lastmod)
;