Actually, Postgres does have a JSON datatype. The only extra thing it does is ensure the data is valid JSON, but it is still technically incorrect to say that there is no JSON type.<p><a href="http://www.postgresql.org/docs/devel/static/datatype-json.html" rel="nofollow">http://www.postgresql.org/docs/devel/static/datatype-json.ht...</a>
In related news (here: <a href="https://news.ycombinator.com/item?id=5589593" rel="nofollow">https://news.ycombinator.com/item?id=5589593</a> ), the Opa framework which is based on Node.js just released support for both MongoDB and Postgres from the same application source.
The plv8x project ( <a href="https://github.com/clkao/plv8x" rel="nofollow">https://github.com/clkao/plv8x</a> ) adds convenient JSON and JS/LiveScript selector operators: <a href="https://speakerdeck.com/audreyt/pgrest-node-dot-js-in-the-database" rel="nofollow">https://speakerdeck.com/audreyt/pgrest-node-dot-js-in-the-da...</a> covers the basic usage and the motivation, as well as the ongoing PgREST work ( <a href="https://github.com/clkao/pgrest" rel="nofollow">https://github.com/clkao/pgrest</a> ) that builds a MongoLab REST API server that resides within the database, so they can be queried over a pg connection as well as through HTTP.
Wouldn't it be a lot nicer to just use a schema and serializer? I know that's not really the point of this article, but if you're trying to get a SQL db to operate like a NoSQl store, it feels like a lot of hand-waving.