I need to convert a Json file into a readable csv - json

Question

I got this Perl code that is supposed to read my categories and put them into a csv file. After many tries i finally got it but is only ready 50 of my over 500 categories. Any way to modify this routine to read all my categories.
Here is the Perl file I got from the Bigcommerce forum.
use strict;
use JSON::PP;
open (my $fh, "<", 'categories.json');
my $json_text = <$fh>;
my $perl_scalar = decode_json($json_text);
# Make a list of ids to names, so that I can build a content path for Neto category CSV
my $id;
foreach my $element (#$perl_scalar)
{
$id->{$element->{id}}=$element->{name};
}
# Actually print out the CSV content, in Neto's required format.
print "content type,content path,name,description 1,description 2,sort order,seo meta description,seo page title,seo meta keywords\n";
foreach my $element (#$perl_scalar)
{
print "Product Category,";
my $parent_category = $element->{parent_category_list}[0];
if ($parent_category == $element->{id})
{
print ",";
}
else
{
print $id->{$parent_category}, ",";
}
print $element->{name}, ",", $element->{description}, ",,", $element->{sort_order}, ",", $element->{meta_description}, ",,\n";
}
Thanks in advance

Answer 1

There is a pretty fundamental problem with mapping JSON to CSV. JSON is a nested data structure, where CSV isn't. Therefore you'll always have to mess around with converting - how would you colliminate:
{
"data2" : {
"fish" : "paste"
},
"data" : [
{
"somesub" : "somethingelse"
},
{
"somesub" : "anotherthing"
}
]
}
This won't turn into a flat data structure like CSV easily.
If you've some trivial JSON to convert, it's not too hard, but depends entirely on the structure of your JSON file, and how you want to map things.
For a trivial example:
use strict;
use warnings;
use JSON;
use Data::Dumper;
local $/;
my $data = from_json(<DATA>);
print Dumper $data;
my #columns = qw ( col1 col2 col3 );
print join( ",", "key", #columns ), "\n";
foreach my $key ( sort keys %$data ) {
print join( ",", $key, #{ $data->{$key} }{#columns} ), "\n";
}
__DATA__
{
"1" :
{
"col1" : "value1",
"col2" : "value2",
"col3" : "value3"
},
"2" : {
"col1" : "value4",
"col2" : "value5",
"col3" : "value6"
}
}
For a more complex example - it may be appropriate to use Text::CSV - but it depends rather what's in your JSON content - the simplistic join approach above doesn't cope with line feeds, embedded quotes or commas within the text. So it might be better to use Text::CSV:
#!/usr/bin/env perl
use strict;
use warnings;
use JSON;
use Text::CSV;
use Data::Dumper;
local $/;
my $data = from_json ( <DATA> );
print Dumper $data;
my $csv = Text::CSV -> new ( { 'binary' => 1 } );
my #columns = qw ( col1 col2 col3 );
$csv -> column_names ( #columns );
foreach my $key ( sort keys %$data ) {
$csv -> print_hr ( \*STDOUT, $data->{$key} );
print "\n";
}
foreach my $key ( sort keys %$data ) {
my $row = [ $key, #{$data->{$key}}{#columns} ];
$csv -> print ( \*STDOUT, $row );
print "\n";
}
This uses the same __DATA__ block as above, and also runs twice - once with using 'column headings' to print - which works provided you don't want to preserve the "key" field, and the second which assembles an array reference to print.