What kind of data format is this (colon and semicolon separated entries)?

Question

I'm integrating two poorly documented systems, and in the process I've come across a strange data format I haven't seen before. It's stored as plain text in the db with no indication as to what the format is and how to deal with it.

a:17:{s:2:"id";s:27:"145219921F990B11C39E7220000";s:16:"purchase_country";s:2:"no";s:17:"purchase_currency";s:3:"nok";s:6:"locale";s:5:"nb-no";s:6:"status";s:17:"checkout_complete";s:9:"reference";s:27:"145212221F990B11C39E7221000";s:11:"reservation";s:10:"2348226550";s:10:"started_at";s:25:"2014-04-04T10:40:55+02:00";s:12:"completed_at";s:25:"2014-04-02T10:41:11+02:00";s:16:"last_modified_at";s:25:"2014-04-02T10:41:11+02:00";s:10:"expires_at";s:25:"2014-04-16T10:41:11+02:00";s:4:"cart";a:4:{s:25:"total_price_excluding_tax";i:489500;s:16:"total_tax_amount";i:0;s:25:"total_price_including_tax";i:489500;s:5:"items";a:2:{i:0;a:10:{s:9:"reference";s:2:"68";s:4:"name";s:21:"1.OSO SUPER S 200LIT.";s:8:"quantity";i:1;s:10:"unit_price";i:695000;s:8:"tax_rate";i:0;s:13:"discount_rate";i:0;s:4:"type";s:8:"physical";s:25:"total_price_including_tax";i:695500;s:25:"total_price_excluding_tax";i:694000;s:16:"total_tax_amount";i:0;}i:1;a:10:{s:9:"reference";s:2:"68";s:4:"name";s:32:"1.OSO SUPER S 200LIT. (discount)";s:8:"quantity";i:1;s:10:"unit_price";i:-205100;s:8:"tax_rate";i:0;s:13:"discount_rate";i:0;s:4:"type";s:8:"physical";s:25:"total_price_including_tax";i:-205100;s:25:"total_price_excluding_tax";i:-205100;s:16:"total_tax_amount";i:0;}}}s:8:"customer";a:1:{s:4:"type";s:6:"person";}s:16:"shipping_address";a:8:{s:10:"given_name";s:13:"Testperson-no";s:11:"family_name";s:8:"Approved";s:14:"street_address";s:18:"Sæffleberggate 56";s:11:"postal_code";s:4:"0563";s:4:"city";s:4:"OSLO";s:7:"country";s:2:"no";s:5:"email";s:32:"omitted@testdrive.klarna.com";s:5:"phone";s:11:"40 12 34 56";}s:15:"billing_address";a:8:{s:10:"given_name";s:13:"Testperson-no";s:11:"family_name";s:8:"Approved";s:14:"street_address";s:18:"Sæffleberggate 56";s:11:"postal_code";s:4:"0563";s:4:"city";s:4:"OSLO";s:7:"country";s:2:"no";s:5:"email";s:32:"checkout-no@testdrive.klarna.com";s:5:"phone";s:11:"40 12 34 56";}s:7:"options";a:1:{s:31:"allow_separate_shipping_address";b:0;}s:8:"merchant";a:5:{s:2:"id";s:4:"1601";s:9:"terms_uri";s:95:"omitted";s:12:"checkout_uri";s:59:"omitted";s:16:"confirmation_uri";s:220:"omitted";s:8:"push_uri";s:229:"omitted";}} 

An entry consists of colon-separated segments:

• A single char type tag (array, object, int, decimal, bool, string)
• A number that says how long the value is in characters, bytes, elements (in case of arrays) or key-value pairs (in case of objs), which seems completely useless given that this is a textual format that requires me to parse the length segment anyway. This isn't present for integers and decimals.
• Value of the field
• Key-value pairs seem to be a flat list of an even number of elements. They also seem to be using arrays as objects as well (see example).
• A ; terminator, which seems not to be necessary for objects and arrays, just to make parsing more tedious.

Now, parsing this thing is reasonably easy, but I'm constantly being surprised by new data types and their weird syntax and I'm not sure that I've covered all the edge cases with the few data samples I've analyzed. Is anyone familiar with this format?

Answer 1

Looks like PHP serialization. See: http://www.phpinternalsbook.com/classes_objects/serialization.html