How to skip a directory in awk?

Question

Say I have the following structure of files and directories:

$ tree
.
├── a
├── b
└── dir
    └── c

1 directory, 3 files

That is, two files a and b together with a dir dir, where another file c stands.

I want to process all the files with awk (GNU Awk 4.1.1, exactly), so I do something like this:

$ gawk '{print FILENAME; nextfile}' * */*
a
b
awk: cmd. line:1: warning: command line argument `dir' is a directory: skipped
dir/c

All is fine but the * also expands to the directory dir and awk tries to process it.

So I wonder: is there any native way awk can check if the given element is a file or not and, if so, skip it? That is, without using system() for it.

I made it work by calling the external system in BEGINFILE:

$ gawk 'BEGINFILE{print FILENAME; if (system(" [ ! -d " FILENAME " ]")) {print FILENAME, "is a dir, skipping"; nextfile}} ENDFILE{print FILENAME, FNR}' * */*
a
a 10
a.wk
a.wk 3
b
b 10
dir
dir is a dir, skipping
dir/c
dir/c 10

_{Note also the fact that if (system(" [ ! -d " FILENAME " ]")) {print FILENAME, "is a dir, skipping"; nextfile} works counter intuitively: it should return 1 when true, but it returns the exit code.}

I read in A.5 Extensions in gawk Not in POSIX awk:

• Directories on the command line produce a warning and are skipped (see Command-line directories)

And then the linked page says:

4.11 Directories on the Command Line

According to the POSIX standard, files named on the awk command line must be text files; it is a fatal error if they are not. Most versions of awk treat a directory on the command line as a fatal error.

By default, gawk produces a warning for a directory on the command line, but otherwise ignores it. This makes it easier to use shell wildcards with your awk program:

$ gawk -f whizprog.awk *        Directories could kill this program

If either of the --posix or --traditional options is given, then gawk reverts to treating a directory on the command line as a fatal error.

See Extension Sample Readdir, for a way to treat directories as usable data from an awk program.

And in fact it is the case: the same command as before with --posix fails:

$ gawk --posix 'BEGINFILE{print FILENAME; if (system(" [ ! -d " FILENAME " ]")) {print FILENAME, "is a dir, skipping"; nextfile}} ENDFILE{print FILENAME, NR}' * */*
gawk: cmd. line:1: fatal: cannot open file `dir' for reading (Is a directory)

I checked the 16.7.6 Reading Directories section that is linked above and they talk about readdir:

The readdir extension adds an input parser for directories. The usage is as follows:

@load "readdir"

But I am not sure neither how to call it nor how to use it from the command line.

Answer 1

If you wanted to safeguard your script from other people mistakenly passing a directory (or anything else that's not a readable text file) to it, you could do this:

$ ls -F tmp
bar  dir/  foo

$ cat tmp/foo
line 1

$ cat tmp/bar
line 1
line 2

$ cat tmp/dir
cat: tmp/dir: Is a directory

$ cat tst.awk
BEGIN {
    for (i=1;i<ARGC;i++) {
        if ( (getline line < ARGV[i]) <= 0 ) {
            print "Skipping:", ARGV[i], ERRNO
            delete ARGV[i]
        }
        close(ARGV[i])
    }
}
{ print FILENAME, $0 }

$ awk -f tst.awk tmp/*
Skipping: tmp/dir Is a directory
tmp/bar line 1
tmp/bar line 2
tmp/foo line 1

$ awk --posix -f tst.awk tmp/*
Skipping: tmp/dir
tmp/bar line 1
tmp/bar line 2
tmp/foo line 1

Per POSIX getline returns -1 if/when it fails trying to retrieve a record from a file (e.g. unreadable file or file does not exist or file is a directory), you just need GNU awk to tell you which of those failures it was by the value of ERRNO if you care.