Date: Mon, 20 Jul 2020 13:18:55 +0200
From: Hans Ginzel <hans@matfyz.cz>
To: psycopg@lists.postgresql.org
Subject: Re: register_adapter Json with custom JSONEncoder
Message-ID: <20200720111854.GA63808@artax.karlin.mff.cuni.cz>
Mail-Followup-To: psycopg@lists.postgresql.org
References: <20200713154907.GH81733@artax.karlin.mff.cuni.cz>
 <CA+mi_8bPLhcgA3KgyZ-g+=oSkUBjaB-JoZD5oAKVjPqbm8iiog@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Disposition: inline
In-Reply-To: 
 <CA+mi_8bPLhcgA3KgyZ-g+=oSkUBjaB-JoZD5oAKVjPqbm8iiog@mail.gmail.com>
User-Agent: Mutt/1.5.21 (2010-09-15)
Precedence: bulk

On Mon, Jul 13, 2020 at 05:15:19PM +0100, Daniele Varrazzo wrote:
>You can use a partial function for instance, of the like of: `lambda
>obj: Json(obj, dumps=mydumps)`.

Thank you, Daniele.
I have (succesfully) tried

from psycopg2.extras import Json
from json import JSONEncoder
from bson import ObjectId

class JSONEncoder(JSONEncoder):
     def default(self, o):
         if isinstance(o, ObjectId):
             return str(o)
         return supper().default(self, o)
jsonEncoder = JSONEncoder()
#psycopg2.extensions.register_adapter(dict, Json)
psycopg2.extensions.register_adapter(dict, lambda o: Json(o, dumps=jsonEncoder.encode))

and variants like

import json
psycopg2.extensions.register_adapter(dict, lambda o: Json(o, dumps = lambda oo: json.dumps(oo, default=str)))

or

from bson import json_util
psycopg2.extensions.register_adapter(dict, lambda o: Json(o, dumps=json_util.dumps))

I choosed the first one because of cached Encoder and cleanest way.

Best regards,
Hans