Here's a simple(-minded) way to check a file for the number of commas per line:

perl -lne '$n=tr/,//;$h{$n}++;END{print "$h{$_} rows have $_ commas" f
+or (keys %h)}' some_file.csv
[download]

If that produces only one line of output for a given file, you know the file has the same number of commas in all rows (and it tells you how many commas per row).

If some lines have more and others have fewer, you'll get a breakdown of the variance. It still could be a "valid" CSV file, if lines with extra commas happen to have quotes or escapes (meaning that you really need to use a parsing module like Text::xSV).

Apart from that, even if the CSV data is simple (no quoted/escaped commas) and has the same number of commas on every line, you need to be careful with your use of split() -- this would be wrong:

    split(",")
[download]

You should do it like this instead:

    split( /,/, $_, -1 );
[download]

If you don't do that, split() will ignore "extra" commas at the end of a line -- e.g. this:

   @array = split( /,/, "field1,field2,,field4,field5,,," );
[download]

will fill @array like this:

   ( 'field1', 'field2', undef, 'field4', 'field5' );
[download]

Note how the trailing empty fields are truncated. Please read about split.