lpogoda <lpogodajr292185@COMCAST.NET> replied: > Depending on what created them, human readable files can and do have a > variety of extensions. Conversely, just about any file can be renamed to > anything else without changing it's content. For instance, I've just > finished on a project where 100's of ASCII (in the above sense of human > readable) files were received with an extension of .aws. And due to an > update of some media player or other, my office pc thinks that version 6 SAS > datasets are Macintosh sound files at the moment. In other words, the file > name or extension is not an infallible guide to the file contents. Not by > half. > > So, I read the above as is there any test that will tell you if a > file-that-a-human-can-read-as-if-reading-a-page-of-a-book exists?

Perl has an approximation. In Perl there are a host of operators that look like -X (where 'X' is a single alphabetic character).

-e $file # $file exists -f $file # $file is a 'plain' file -T $file # $file is an ASCII text file

The -T operator reads the first block of the file and counts up the number of control characters and characters with the high bit set. If that total is over 30% of the first block (or if there is a null in the first block) then the file is regarded as other than an ASCII text file, and (-T $file) returns 0 (for false). This could obviously be copied in SAS using recfm=N to scan through the first umpteen bytes and make the same evaluation.

Since you have to read the file to do the -T test, usually you want to check that the file exists first, and abend if it isn't there:

if -e $file && -T $file {other code ...};

HTH, David -- David Cassell, CSC Cassell.David@epa.gov Senior computing specialist mathematical statistician