allow delmiter to be a regex; use regex per column

diff --git a/col_grep b/col_grep
index 22d6464808c94c437883ca34e5c96c8ec4f4960b..9f181cb16998cb04523ace37c2273f9a1c46be68 100755 (executable)
--- a/col_grep
+++ b/col_grep
@@ -102,8 +102,7 @@ if ($options{has_header}) {
 my @field_indexes = map { looks_like_number($_) && $_ > 0 ? $_ - 1 : $_;}
     map {split /,/} @{$options{field}};
 my %field_indexes;
-@field_indexes{@field_indexes} = @field_indexes;
-@field_indexes = values %field_indexes;
+@field_indexes = map {++$field_indexes{$_} > 1 ? ():$_} @field_indexes;
 
 if (grep {not /-?\d+/} @field_indexes and not $options{has_header}) {
     push @USAGE_ERRORS,"Invalid field index(es)";
@@ -120,7 +119,13 @@ if (not @ARGV) {
     push @ARGV,undef;
 }
 my %compiled_regexes;
-my $csv = Text::CSV->new({sep_char=>$options{delimiter}});
+my $csv;
+if (length($options{delimiter}) > 1) { # delimiter is a regex, then
+    #csv will be undef
+} else {
+    $csv = Text::CSV->new({sep_char=>$options{delimiter}}) or
+        die "Unable to start Text::CSV";
+} 
 FILE: for my $file (@ARGV) {
     my $fh;
     my $headers_updated = 0;
@@ -134,8 +139,13 @@ FILE: for my $file (@ARGV) {
  LINE: while (<$fh>) {
         my $chomped = chomp;
         my $line = $_;
-        die "Unable to parse line $. of $file: ".$csv->error_diag() unless $csv->parse($_);
-        my @fields = $csv->fields();
+        my @fields;
+        if (defined $csv) {
+            die "Unable to parse line $. of $file: ".$csv->error_diag() unless $csv->parse($_);
+            @fields = $csv->fields();
+        } else {
+            @fields = split /$options{delimiter}/o,$_;
+        }
         # skip lines which don't have enough fields
         if ($options{has_header} and not @header) {
             @header = @fields;
@@ -145,7 +155,7 @@ FILE: for my $file (@ARGV) {
             next LINE;
         }
         if ($options{has_header} and not $headers_updated) {
-            $headers_updated = 0;
+            $headers_updated = 1;
             if (@header < @fields) {
                 @header{@header} = 1..@fields;
             }
@@ -154,7 +164,7 @@ FILE: for my $file (@ARGV) {
                 push @new_indexes,$index and next if $index =~ /^-?\d+$/;
                 if (not exists $header{$index}) {
                     use Data::Dumper;
-                    print STDERR Dumper(\%header);
+                    print STDERR Data::Dumper->Dump([\%header],[qw(*header)]);
                     print STDERR "Invalid header $index\n";
                     exit 1;
                 } else {
@@ -162,16 +172,25 @@ FILE: for my $file (@ARGV) {
                 }
             }
             @field_indexes = @new_indexes;
+            print STDERR Data::Dumper->Dump([\@field_indexes],[qw(*field_indexes)]) if $DEBUG;
         }
         next LINE if grep {not defined $_} @fields[@field_indexes];
+        my $i = -1;
     REGEX: for my $regex (@{$options{regexp}}) {
-        FIELDS: for my $field (@fields[@field_indexes]) {
-                if (length $regex) {
+            $i++;
+            if (length $regex) {
+                my @fields_to_examine = map {$fields[$_]} @field_indexes;
+                if (@{$options{regexp}} > 1) {
+                    @fields_to_examine = $fields_to_examine[$i];
+                }
+            FIELDS: for my $field (@fields_to_examine) {
                     if (not exists $compiled_regexes{$regex}) {
                         $compiled_regexes{$regex} = qr/$regex/;
                     }
-                    $field =~ $compiled_regexes{$regex} or next LINE;
+                    print STDERR "regex: $regex field: $field\n" if $DEBUG;
+                    $field =~ $compiled_regexes{$regex} and next REGEX;
                 }
+                next LINE;
             }
         }
         print $line;