NAME
Text::xSV::Slurp - Convert xSV data to common data shapes.
VERSION
Version 0.20
SYNOPSIS
Text::xSV::Slurp
converts between xSV data and a variety of nested data shapes, allowing both column and row filtering using user defined functions.
This brief example creates an array of hashes from a file, where each array record corresponds to a line of the file, and each line is represented as a hash of header-to-value pairs.
use Text::xSV::Slurp 'xsv_slurp';
my $aoh = xsv_slurp( 'foo.csv' );
## if foo.csv contains:
##
## uid,name
## 342,tim
## 939,danboo
##
## then $aoh contains:
##
## [
## { uid => '342', name => 'tim' },
## { uid => '939', name => 'danboo' },
## ]
FUNCTIONS
xsv_slurp()
xsv_slurp()
converts an xSV data source to one of a variety of nested data shapes. It allows both column and row filtering using user defined functions.
Option summary:
file
- file name to be openedhandle
- file handle to be iteratedstring
- string to be parsedshape
- target data structure (aoa
,aoh
,hoa
orhoh
)col_grep
- skip a subset of columns based on user callbackrow_grep
- skip a subset of rows based on user callbackkey
- xSV string or ARRAY used to build the keys of thehoh
shapetext_csv
- option hash for Text::CSV/Text::CSV_XS constructor
The file
, handle
and string
options are mutually exclusive. Only one source parameter may be passed in each call to xsv_slurp()
, otherwise a fatal exception will be raised.
The source can also be provided implicitly, without the associated key, and the source type will be guessed by examining the first item in the option list. If the item is a reference type, it is treated as a handle
source. If the item contains a newline or carriage return, it is treated as a string
source. If the item passes none of the prior tests, it is treated as a file
source.
## implicit C<handle> source
my $aoa = xsv_slurp( \*STDIN, shape => 'aoa' );
## implicit C<string> source
my $aoh = xsv_slurp( "h1,h2\n" . "d1,d2\n" );
## implicit C<file> source
my $aoh = xsv_slurp( 'foo.csv' );
The shape
parameter supports values of aoa
, aoh
, hoa
or hoh
. The default shape is aoh
. Each shape affects certain parameters differently (see below).
The text_csv
option can be used to control Text::CSV/Text::CSV_XS parsing. The given HASH reference is passed to the Text::CSV constructor. If the text_csv
option is undefined, the default Text::CSV constructor is called. For example, to change the separator to a colon, you could do the following:
my $aoh = xsv_slurp( file => 'foo.csv',
text_csv => { sep_char => ':' } );
aoa
example input:
h1,h2,h3
l,m,n
p,q,r
example data structure:
[
[ qw/ h1 h2 h3 / ],
[ qw/ l m n / ],
[ qw/ p q r / ],
]
shape specifics:
col_grep
- passed an ARRAY reference of indexes, should return a list of indexes to be includedrow_grep
- passed an ARRAY reference of values, should return true or false whether the row should be included or not
full example:
## - convert xSV example to an array of arrays
## - include only rows containing values matching /[nr]/
## - include only the first and last columns
my $aoa = xsv_slurp( string => $xsv_data,
shape => 'aoa',
col_grep => sub { return @( shift() }[0,-1] },
row_grep => sub { return grep /[nr]/, @{ $_[0] } },
);
## $aoa contains:
##
## [
## [ 'l', 'n' ],
## [ 'p', 'r' ],
## ]
aoh
example input:
h1,h2,h3
l,m,n
p,q,r
example data structure:
[
{ h1 => 'l', h2 => 'm', h3 => 'n' },
{ h1 => 'p', h2 => 'q', h3 => 'r' },
]
shape specifics:
col_grep
- passed an ARRAY reference of column names, should return a list of column names to be includedrow_grep
- passed a HASH reference of column name / value pairs, should return true or false whether the row should be included or not
full example:
## - convert xSV example to an array of hashes
## - include only rows containing values matching /n/
## - include only the h3 column
my $aoh = xsv_slurp( string => $xsv_data,
shape => 'aoh',
col_grep => sub { return 'h3' },
row_grep => sub { return grep /n/, values %{ $_[0] } },
);
## $aoh contains:
##
## [
## { h3 => 'n' },
## ]
hoa
example input:
h1,h2,h3
l,m,n
p,q,r
example data structure:
{
h1 => [ qw/ l p / ],
h2 => [ qw/ m q / ],
h3 => [ qw/ n r / ],
}
shape specifics:
col_grep
- passed an ARRAY reference of column names, should return a list of column names to be includedrow_grep
- passed a HASH reference of column name / value pairs, should return true or false whether the row should be included or not
full example:
## - convert xSV example to a hash of arrays
## - include only rows containing values matching /n/
## - include only the h3 column
my $hoa = xsv_slurp( string => $xsv_data,
shape => 'hoa',
col_grep => sub { return 'h3' },
row_grep => sub { return grep /n/, values %{ $_[0] } },
);
## $hoa contains:
##
## {
## h3 => [ qw/ n r / ],
## }
hoh
example input:
h1,h2,h3
l,m,n
p,q,r
example data structure (assuming a key
of 'h2,h3'
):
{
m => { n => { h1 => 'l' } },
q => { r => { h1 => 'p' } },
}
shape specifics:
key
- an xSV string or ARRAY specifying the indexing column namescol_grep
- passed an ARRAY reference of column names, should return a list of column names to be includedrow_grep
- passed a HASH reference of column name / value pairs, should return true or false whether the row should be included or noton_collide
- specify how key collisions should be handled (see "HoH collision handlers")
full example:
## - convert xSV example to a hash of hashes
## - index using h1 values
## - include only rows containing values matching /n/
## - include only the h3 column
my $hoh = xsv_slurp( string => $xsv_data,
shape => 'hoh',
key => 'h1',
col_grep => sub { return 'h3' },
row_grep => sub { return grep /n/, values %{ $_[0] } },
);
## $hoh contains:
##
## {
## l => { h3 => 'n' },
## p => { h3 => 'r' },
## }
HoH storage handlers
Using the hoh
shape can result in non-unique key
combinations. The default action is to simply assign the values to the given slot as they are encountered, resulting in any prior values being lost.
For example, using h1,h2
as the indexing key with the default collision handler:
$xsv_data = <<EOXSV;
h1,h2,h3
1,2,3
1,2,5
EOXSV
$hoh = xsv_slurp( string => $xsv_data,
shape => 'hoh',
key => 'h1,h2'
);
would result in the initial value in the h3
column being lost. The resulting data structure would only record the 5
value:
{
1 => { 2 => { h3 => 5 } }, ## 3 sir!
}
Typically this is not very useful. The user probably wanted to aggregate the values in some way. This is where the on_store
and on_collide
handlers come in, allowing the caller to specify how these assignments should be handled.
The on_store
handler is called for each assignment action, while the on_collide
handler is only called when an actual collision occurs (i.e., the nested value path for the current line is the same as a prior line).
If instead we wanted to push the values onto an array, we could use the built-in push
handler for the on_store
event as follows:
$hoh = xsv_slurp( string => $xsv_data,
shape => 'hoh',
key => 'h1,h2',
on_store => 'push',
);
the resulting HoH
, using the same data as above, would instead look like:
{
1 => { 2 => { h3 => [3,5] } }, ## 3 sir!
}
Or if we wanted to sum the values we could us the sum
handler for the on_collide
event:
$hoh = xsv_slurp( string => $xsv_data,
shape => 'hoh',
key => 'h1,h2',
on_collide => 'sum',
);
resulting in the summation of the values:
{
1 => { 2 => { h3 => 8 } },
}
builtin on_store
handlers
A number of builtin on_store
handlers are provided and can be specified by name.
The example data structures below use the following data.
h1,h2,h3
1,2,3
1,2,5
count
Count the times a key occurs.
{ 1 => { 2 => { h3 => 2 } } }
frequency
Create a frequency count of values.
{ 1 => { 2 => { h3 => { 3 => 1, 5 => 1 } } } }
push
push
values onto an array *always*.
{ 1 => { 2 => { h3 => [ 3, 5 ] } } }
unshift
unshift
values onto an array *always*.
{ 1 => { 2 => { h3 => [ 5, 3 ] } } }
builtin on_collide
handlers
A number of builtin on_collide
handlers are provided and can be specified by name.
The example data structures below use the following data.
h1,h2,h3
1,2,3
1,2,5
sum
Sum the values.
{ 1 => { 2 => { h3 => 8 } } }
average
Average the values.
{ 1 => { 2 => { h3 => 4 } } }
push
push
values onto an array *only on colliding*.
{ 1 => { 2 => { h3 => [ 3, 5 ] } } }
unshift
unshift
values onto an array *only on colliding*.
{ 1 => { 2 => { h3 => [ 5, 3 ] } } }
die
Carp::confess if a collision occurs.
Error: key collision in HoH construction (key-value path was: { 'h1' => '1' }, { 'h2' => '2' })
warn
Carp::cluck if a collision occurs.
Warning: key collision in HoH construction (key-value path was: { 'h1' => '1' }, { 'h2' => '2' })
AUTHOR
Dan Boorstein, <dan at boorstein.net>
TODO
add weighted-average collide keys and tests
document hoh 'on_store/on_collide' custom keys
add a recipes/examples section to cover grep and on_collide examples
BUGS
Please report any bugs or feature requests to bug-text-xsv-slurp at rt.cpan.org
, or through the web interface at http://rt.cpan.org/NoAuth/ReportBug.html?Queue=Text-xSV-Slurp. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.
SUPPORT
You can find documentation for this module with the perldoc command.
perldoc Text::xSV::Slurp
You can also look for information at:
RT: CPAN's request tracker
AnnoCPAN: Annotated CPAN documentation
CPAN Ratings
Search CPAN
ACKNOWLEDGEMENTS
COPYRIGHT & LICENSE
Copyright 2009 Dan Boorstein.
This program is free software; you can redistribute it and/or modify it under the terms of either: the GNU General Public License as published by the Free Software Foundation; or the Artistic License.
See http://dev.perl.org/licenses/ for more information.