FYI, the regexp search in the sparse attributes: http://bazaar.launchpad.net/~camptocamp/openerp-product-attributes/6.1-backport-lot-custom-attributes-lep/view/head:/production_lot_custom_attributes/lot.py#L34
The MP is here: https://code.launchpad.net/~camptocamp/openerp-product-attributes/6.1-backport-lot-custom-attributes-lep/+merge/196538 Thanks On Tue, Dec 10, 2013 at 2:18 PM, Leonardo Pistone < [email protected]> wrote: > Thanks for yor answer Raphael. > > On the long run, I agree with you that using a native Postgresql field > seems like the obvious choice. > > For now, searching in the text-based json fields seems something to > implement with a Regexp. Otherwise, maybe with postgresql 9.2 we can use > the native json methods to search in the text field, but I'm not sure. > > Best, Leo > > > On Tue, Dec 10, 2013 at 2:07 PM, Raphael Valyi <[email protected]> wrote: > >> On Tue, Dec 10, 2013 at 10:54 AM, Leonardo Pistone < >> [email protected]> wrote: >> >>> Hi all, >>> >>> Has anyone worked on a search() method for JSON-serialized sparse fields? >>> >>> In theory, one could use postgresql's JSON features introduced in 9.2, >>> or failing that, some (dirty) regexp to build the search query. >>> >>> Any ideas or suggestions? >>> >>> I've seen Therp's unserialize_field, but I like to stay serialized :) >>> >>> Thanks! >>> >> >> >> Hello Leonardo, >> >> For flexibility I would favor evolving the sparse field feature from a >> text field to a native PostgreSQL json field. I mean I say this because I >> think it's better than moving to hstore and getting the search feature from >> hstore. >> I have some experience extending custom_product_attributes into a product >> configurator (though funding was dropped 2 years ago) and I can tell you >> nesting is definitely a must have in the big picture. >> See http://www.craigkerstiens.com/2013/07/03/hstore-vs-json/ >> >> Then yes, I would naively go for building a regexp to add some search >> support upon a full text indexed field. >> >> Also notice that if you admit a heavier solution, I built an Apache SolR >> connector providing much more advanced search features >> https://github.com/akretion/solerp/tree/master/solerp >> The real case I've been working has nearly 1 millions of products indexed >> into Apache SolR, and it's indexing thousands of attributes that use the >> product_custom_attributes and hence sparse fields NOSQL power. >> >> I'm curious what other think about this too. >> >> Regards. >> >> -- >> Raphaël Valyi >> Founder and consultant >> http://twitter.com/rvalyi <http://twitter.com/#!/rvalyi> >> +55 21 2516 2954 >> www.akretion.com >> >> >> >
_______________________________________________ Mailing list: https://launchpad.net/~openerp-expert-framework Post to : [email protected] Unsubscribe : https://launchpad.net/~openerp-expert-framework More help : https://help.launchpad.net/ListHelp
