Issue 31466: No easy way to change float formatting when subclassing encoder.JSONEncoder (original) (raw)

Created on 2017-09-14 09:48 by qpeter, last changed 2022-04-11 14:58 by admin. This issue is now closed.

Messages (4)
msg302162 - (view)	Author: Quentin Peter (qpeter) *	Date: 2017-09-14 09:48
I want to output floats in ENG format. Working with distance in micrometers, it is a bit annoying to see: 2.5e-5 .0003 instead of 25e-6 300e-6 The solution I found was to redefine `iterencode` but that doesn't feel right. I would like to see something similar to the "default" function that is easely modified.
msg302306 - (view)	Author: Éric Araujo (eric.araujo) *	Date: 2017-09-16 00:44
Hello and thanks for the patch! If you don’t get feedback, you could try the python-ideas mailing list to see if other people have the same need and if there’s already a solution for this. Serhiy, would you mind refreshing my memory about json development? I’m not sure whether it’s still supposed to follow the development of simplejson, or if it’s effectively a fork now.
msg387891 - (view)	Author: Inada Naoki (methane) *	Date: 2021-03-02 01:09
May I close this? There is no discussion in python-dev/python-ideas. I feel that this proposal is too minor. Stdlib shouldn't add option for such minor use case.
msg389993 - (view)	Author: Samuel Freilich (sfreilich) *	Date: 2021-04-01 15:16
I think the less-minor issue, of which this is a small subset, is that JSONEncoder doesn't allow changing the behavior for default-serializable types at all. That means you can't choose to lose less information in round-trip serialization/deserialization, if that's what you want (e.g. there's no way to round-trip serialize a tuple with JSONEncoder, though it's trivial to do that for a set).