Text::CSV::LibCSV

Uses a wrapper that can stream data/records from a csv file using libcsv.

Options

Delimiter (.new(:delimiter<>))

LibCSV allows a char to be specified for the delimiter, if you provide a multicharacter string only the first ordinal will be used.

Quote (.new(:quote<>))

LibCSV allows a char to be specified for the quoting char, if you provide a multicharacter string only the first ordinal will be used.

Preserve Whitespace (.new(:preserve-ws))

You can choose to not auto-trim tabs and spaces from data by passing this option along

Parser Options (.new(:parser-options(int32)))

See your local libcsv documentation for what is acceptable here, out of the box is provided the structure:

CSV-STRICT        => 1
CSV-REPALL-NL     => 2
CSV-STRICT-FINI   => 4
CSV-APPEND-NULL   => 8
CSV-EMPTY-IS-NULL => 16

You can build your own options with the enum (example: CSV-STRICT +| CSV-EMPTY-IS-NULL) or the convenience method: Text::CSV::LibCSV.build-options(:CSV-STRICT, :CSV-EMPTY-IS-NULL)

Has Headers (.new(:has-headers) | .read-file(:has-headers))

When provided to .new this option will set the default parsing mode to assuming there are headers in the CSV file. This option can be overridden by passing this option to .read-file. This option will also cause a hash containing header => <value> pairs rather than array to be returned from parsing

Auto Decode (.new(:auto-decode))

This will auto-decode the data records that are emitted from libcsv. If this option is left empty then the user will receive Buf types back

Usage

OO interface

use Text::CSV::LibCSV;

my $parser-options = CSV-STRICT +| CSV-EMPTY-IS-NULL;
my Text::CSV::LibCSV $parser .=new(:$parser-options, :auto-decode('utf8'), :has-headers);

my @lines = $parser.read-file('path-to-file');

# @lines = [ { ... }, { ... }, ... ];

Procedural

use Text::CSV::LibCSV :csv-read-file;

my @lines = csv-read-file('path-to-file', :$parser-options, :auto-decode('utf8'), :has-headers);

# @lines = [ { ... }, { ... }, ... ];

Streaming

Sometimes your CSV files are large or simply want to act as the record/data is available. This is possible with the following pattern:

use Text::CSV::LibCSV;

my Supplier $on-data   .=new;
my Supplier $on-record .=new;

$on-data.Supply.tap( -> $data {
  # depending on your options $data contains a Buf or Buf.decode()
});

$on-record.Supply.tap( -> $record {
  # depending on your options $record contains a hash or array
});

my $parser-options = CSV-STRICT +| CSV-EMPTY-IS-NULL;
my Text::CSV::LibCSV $parser .=new(:$parser-options, :auto-decode('utf8'), :has-headers);

$parser.read-file('path-to-file', :$on-data, :$on-record, :!return-all);