view release on metacpan or  search on metacpan

lib/App/CSVUtils/csv_munge_rows.pm  view on Meta::CPAN

=head2 csv_munge_rows

Usage:

 csv_munge_rows(%args) -> [$status_code, $reason, $payload, \%result_meta]

Modify CSV data rows using Perl code.

Examples:

=over

=item * Modify two fields in a CSV:

 csv_munge_rows(
     input_filename => "file.csv",
   eval => "\$_->{field1} *= 2; \$_->{field2} =~ s/foo/bar/",
   hash => 1
 );

=back

Perl code (-e) will be called for each row (excluding the header row) and C<$_>
will contain the row (arrayref, or hashref if C<-H> is specified). The Perl code
is expected to modify it.

Aside from C<$_>, C<$main::row> will contain the current row array.
C<$main::rownum> contains the row number (2 means the first data row).
C<$main::csv> is the L<Text::CSV_XS> object. C<$main::fields_idx> is also
available for additional information.

The modified C<$_> will be rendered back to CSV row.

You cannot add new fields using this utility. To do so, use
L<csv-add-fields>. You also cannot delete fields (they just become empty
string if you delete the field in the eval code). To delete fields, use
L<csv-delete-fields>.

Note that you can also munge a single field using L<csv-munge-field>.

This function is not exported.

Arguments ('*' denotes required arguments):

=over 4

=item * B<eval>* => I<str|code>

Perl code.

=item * B<hash> => I<bool>

Provide row in $_ as hashref instead of arrayref.

=item * B<inplace> => I<true>

Output to the same file as input.

Normally, you output to a different file than input. If you try to output to the
same file (C<-o INPUT.csv -O>) you will clobber the input file; thus the utility
prevents you from doing it. However, with this C<--inplace> option, you can
output to the same file. Like perl's C<-i> option, this will first output to a
temporary file in the same directory as the input file then rename to the final
file at the end. You cannot specify output file (C<-o>) when using this option,
but you can specify backup extension with C<-b> option.

Some caveats:

=over

=item * if input file is a symbolic link, it will be replaced with a regular file;

=item * renaming (implemented using C<rename()>) can fail if input filename is too long;

=item * value specified in C<-b> is currently not checked for acceptable characters;

=item * things can also fail if permissions are restrictive;

=back

=item * B<inplace_backup_ext> => I<str> (default: "")

Extension to add for backup of input file.

In inplace mode (C<--inplace>), if this option is set to a non-empty string, will
rename the input file using this extension as a backup. The old existing backup
will be overwritten, if any.

=item * B<input_escape_char> => I<str>

Specify character to escape value in field in input CSV, will be passed to Text::CSV_XS.

Defaults to C<\\> (backslash). Overrides C<--input-tsv> option.

=item * B<input_filename> => I<filename> (default: "-")

Input CSV file.

Use C<-> to read from stdin.

Encoding of input file is assumed to be UTF-8.

=item * B<input_header> => I<bool> (default: 1)

Specify whether input CSV has a header row.

By default, the first row of the input CSV will be assumed to contain field
names (and the second row contains the first data row). When you declare that
input CSV does not have header row (C<--no-input-header>), the first row of the
CSV is assumed to contain the first data row. Fields will be named C<field1>,
C<field2>, and so on.

=item * B<input_quote_char> => I<str>

Specify field quote character in input CSV, will be passed to Text::CSV_XS.

Defaults to C<"> (double quote). Overrides C<--input-tsv> option.

=item * B<input_sep_char> => I<str>