Changeset 38797 for main/trunk

Timestamp:

2024-02-27T14:43:48+13:00 (4 months ago)

Author:

kjdon

Message:

added option '-ignore_field' - if there is a column with this name in the spreadsheet, and a line is non empty for this column, then ignore the line.

File:

• main/trunk/greenstone2/perllib/plugins/CSVPlugin.pm (modified) (6 diffs)

main/trunk/greenstone2/perllib/plugins/CSVPlugin.pm

@@ -74 +74 @@
     'type' => "flag",
     'reqd' => "no"},
+       { 'name' => "ignore_field",
+    'desc' => "{CSVPlugin.ignore_field}",
+    'type' => "string",
+    'reqd' => "no"},
 
 
@@ -109 +113 @@
     if ((defined $self->{'use_namespace_for_field_names'}) && ($self->{'use_namespace_for_field_names'} =~ m/^\s*$/)) {
     $self->{'use_namespace_for_field_names'} = undef;
+    }
+    if ((defined $self->{'ignore_field'}) && ($self->{'ignore_field'} =~ m/^\s*$/)) {
+    $self->{'ignore_field'} = undef;
     }
 
@@ -226 +233 @@
     my $found_filename_field = 0;
     my $filename_field = $self->{'filename_field'};
+    my $ignore_field = $self->{'ignore_field'};
+    my $ignore_col;
     for (my $i = 0; $i < scalar(@csv_file_fields); $i++) {
     # Remove any spaces from the field names, and surrounding quotes too
@@ -232 +241 @@
     $csv_file_fields[$i] =~ s/"$//;
 
+
     if ($self->{'use_namespace_for_field_names'}) {
         $csv_file_fields[$i] =  $self->{'use_namespace_for_field_names'}. "." . $csv_file_fields[$i];
@@ -237 +247 @@
     if ($csv_file_fields[$i] eq  $filename_field) {
         $found_filename_field = 1;
-    } 
+    }
+    if ($ignore_field && $csv_file_fields[$i] eq $ignore_field) {
+        $ignore_col = $i;
+    }
+
     }
 
@@ -246 +260 @@
     }
 
+
+
     my $count = 0;
     while (my $csv_line = $csv->getline($CSV_FILE)) {
     my @md_vals = @$csv_line;
+    
+    if (defined $ignore_col && $md_vals[$ignore_col] ne "") {
+        # ignore this line
+        print STDERR "ignoring line ".join(",", @md_vals)."\n";
+        next;
+    }
     $count++;
-
+    
     # Build a hash of metadata name to metadata value for this line
     my %csv_line_metadata;