Pages

SyntaxHighlighter

Sunday, May 10, 2015

Delete External Files using Wildcards

In the age of cloud computing you may not know the operating system (OS) that SAS is using. As a result, it is best to use SAS functions and statements that are OS independent.

SAS functions are used in the below code to delete one or multiple external files via DOS like wildcard characters. This means that the question mark (?) represents a single character while an asterisk (*) is used for one or more characters. However, those DOS wildcard symbols do not directly translate, so Perl Regular Expression (PRX) functions included in the base SAS language are used.

The FINDC() function is used to find the last slash (\ used by Windows) or backslash (/ used by Linux/UNIX). FINDC() will look for either the / or \ character starting from the end of the string via the -9999 parameter. Once the file portion of the &filename macro variable has been parsed, the TRANWRD function is used to substitute the correct Perl Regular Expression (PRX) characters. PRX reserves the period (.) to represent character so it needs to be escaped (\) as in \. to find a literal period. The entire regular expression is enclosed between the /^ (match beginning of string) to /i (case insensitive)

From there the path or folder is opened using the DOPEN function. The number of entries in the folder are determined by the DNUM function and then the filenames in the folder are read using the DREAD function. After that, the PRXMATCH function is used to determine if the file matches the supplied pattern. Files that match the pattern are then verified with FEXIST and deleted using the FDELETE function.

Here are a few examples of how this can be used followed by the source code:

%deletefiles( filename = c:\temp\xyz.csv ) /* single file */
%deletefiles( filename = c:\temp\xy?.csv ) /* ? for single character */
%deletefiles( filename = c:\temp\x*.csv ) /* files with X ends with .CSV */

%macro deletefiles( filename = ) ;
  %if %sysevalf( %superq( filename ) =,  boolean ) %then %do ;
    %put %str(E)RROR: filename was not supplied.;
    %return ;
  %end ;

  data _null_ ;
    length 
      path file name fullfilename prxstring $256 
      dir filep $8 
      rc did i cnt 4 ;

    path      = substr( "&filename.", 1, findc( "&filename", "\/", -9999 ) ) ;
    file      = substr( "&filename.", findc( "&filename", "\/", -9999 ) + 1 ) ;
    prxstring = cats( "/^"
              , tranwrd(tranwrd(tranwrd(file,".", "\."),"*",".*" ),"?",".?")
              , "/i" ) ;
    
    rc = filename( dir, path ) ;
    if rc = 0 then do ;
      did = dopen( dir ) ;
 
      if did > 0 then do ;
        do i = 1 to dnum( did ) ;
          name = dread( did, i ) ;

          if prxmatch( prxstring, name ) then do ;
            fullfilename = cats( path, name ) ;
            rc = filename( filep, fullfilename ) ;
            if rc = 0 and fexist( filep ) then do ;
              rc = fdelete( filep ) ;
              if rc = 0 then put "NOTE: " fullfilename "has been deleted." ;
              cnt + 1 ;
            end ;
            rc = filename( filep, " " ) ;
          end ;
         
        end ;
        rc = dclose( did ) ;
      end ;
      
      rc = filename( dir, " " ) ;
    end ;
    put / "NOTE: " cnt "files were deleted." ;

   run;
%mend ;
 

No comments:

Post a Comment