How to parse list of models with Pydantic

Question

I use Pydantic to model the requests and responses to an API.

I defined a User class:

from pydantic import BaseModel

class User(BaseModel):
  name: str
  age: int

My API returns a list of users which I retrieve with requests and convert into a dict:

users = [{"name": "user1", "age": 15}, {"name": "user2", "age": 28}]

How can I convert this dict to a list of User instances?

My solution for now is

user_list = []
for user in users:
  user_list.append(User(**user))

score 165 · Accepted Answer · edited Aug 02 '23 at 15:45

165

Pydantic V1:

This is now possible using parse_obj_as.

from pydantic import parse_obj_as

users = [
    {"name": "user1", "age": 15}, 
    {"name": "user2", "age": 28}
]

m = parse_obj_as(List[User], users)

Pydantic V2:

Use Type Adapter.

from pydantic import TypeAdapter

users = [
    {"name": "user1", "age": 15}, 
    {"name": "user2", "age": 28}
]

ta = TypeAdapter(List[User])
m = ta.validate_python(users)

edited Aug 02 '23 at 15:45

Yaakov Bressler

9,056
2
45
69

answered Apr 03 '20 at 22:11

David Asaf

1,666
1
6
6

3

Is there a function to do reverse, i.e given List[User] converts List[dict] – Shiv Krishna Jaiswal Aug 16 '21 at 06:41
6

@ShivKrishnaJaiswal, if you're in FastAPI, there's [`fastapi.encoders.jsonable_encoder`](https://fastapi.tiangolo.com/tutorial/encoder/), as in `jsonable_encoder(my_user_list)`. – LeoRochael Jan 21 '22 at 17:45
1

I got the unhappy `__root__ value is not a valid list (type=type_error.list)` – Jonathan Mugan May 18 '22 at 23:19
2

There is pydantic.json.pydantic_encoder, for those who do not want to install FastAPI for that. Cf https://pydantic-docs.helpmanual.io/usage/dataclasses/#json-dumping – cbenz Nov 10 '22 at 15:07
Pydantic 2.0 update: use `TypeAdapter(List[User]).validate_python(users)`. Moreover, if you have access to the raw JSON `bytes`/`str`, you might also want to delegate the parsing step to Pydantic (skipping the intermediate `dict` representation). e.g. `TypeAdapter(List[User]).validate_json(json_bytes)` – Danilo Gómez Jul 17 '23 at 05:21
parse_obj_as is already deprecated. – Wang Jul 17 '23 at 17:33

score 58 · Answer 2 · edited Apr 13 '21 at 11:48

58

To confirm and expand the previous answer, here is an "official" answer at pydantic-github - All credits to "dmontagu":

The "right" way to do this in pydantic is to make use of "Custom Root Types". You still need to make use of a container model:

class UserList(BaseModel):
    __root__: List[User]

but then the following will work:

UserList.parse_obj([
    {'id': '123', 'signup_ts': '2017-06-01 12:22', 'friends': [1, '2', b'3']},
    {'id': '456', 'signup_ts': '2017-06-02 12:22', 'friends': ['you']},
])

(and will put the values inside the root property).

Unfortunately, I think there is not good serialization support for this yet, so I think when you go to return the results, if you want to return just a list you'll still need to return UserList.root.

I don't think there is currently a unified interface that gets you a serialized/unstructured version of the model that respects the root_model, but if this is what you are looking for, it could be worth building.

edited Apr 13 '21 at 11:48

Timusan

3,255
1
22
27

answered Oct 31 '19 at 10:17

Jiri

691
5
7

1

If you use `.json()` for a model with a custom root type, it'll root the root object (without the `'__root__': `). – SColvin Nov 18 '19 at 13:57
Except when you use `.dict()` it will include `__root__` keys :) – Jeremy Sep 29 '21 at 15:13
To remove `__root__` key, I defined `dict` method which does `return super().dict()['__root__']` – hshib Mar 09 '22 at 22:18
@hshib can u share the snippet of how did it? Thanks in advance. The ugly way I am working with is json.loads(SomeSchema.json()) – Prashant Nair Jun 02 '22 at 18:49

score 19 · Answer 3 · answered Apr 19 '19 at 13:51

19

You can try this

from typing import List
from pydantic import BaseModel

class User(BaseModel):
  name: str
  age: int

class Users(BaseModel):
    users: List[User]

users = [{"name": "user1", "age": 15}, {"name": "user2", "age": 28}]
m = Users(users=users)
print(m.dict())

answered Apr 19 '19 at 13:51

ozcanyarimdunya

2,324
1
18
21

3

Thanks, but that returns an object with a `users` property that contains the list. I will keep it in mind if there is no way to do this, that's nicer! – Nymous Apr 19 '19 at 15:08

score 18 · Answer 4 · answered Aug 27 '19 at 19:53

18

You could consider using a list comprehension along with dict unpacking to the User constructor

user_list = [
  User(**user) for user in users
]

answered Aug 27 '19 at 19:53

svrist

7,042
7
44
67

I like it, pretty clean – lowercase00 Jun 23 '21 at 23:54
simplest and most obvious. should be accepted answer. – rv.kvetch Sep 25 '21 at 13:38
FYI this method is about **2x** faster than the accepted answer. I have no idea why the other one was accepted instead lol. – rv.kvetch Sep 25 '21 at 13:40
@rv.kvetch can you show how you test the speed? – Zaffer Feb 18 '23 at 15:51

score 9 · Answer 5 · answered Oct 31 '19 at 03:41

You can use the __root__ Pydantic keyword:

from typing import List
from pydantic import BaseModel

class User(BaseModel):
  name: str
  age: int

class UserList(BaseModel):
  __root__: List[User]     # ⯇-- __root__

To build the JSON response:

user1 = {"name": "user1", "age": 15}
user2 = {"name": "user2", "age": 28}

user_list = UserList(__root__=[])
user_list.__root__.append(User(**user1))
user_list.__root__.append(User(**user2))

Your API web framework can jsonify user_list to be returned as a JSON array (within the response body).

score 1 · Answer 6 · answered Sep 25 '21 at 08:45

1

I just set in my models.py list of dict like this:

from django.db import models
from pydantic import BaseModel

class CustomList(BaseModel):
    data: list[dict]

answered Sep 25 '21 at 08:45

reza_khalafi

6,230
7
56
82

score 1 · Answer 7 · answered May 03 '23 at 08:23

1

this worked for me using list comprehension:

user_list = [User(**user) for user in users]

answered May 03 '23 at 08:23

Aymen Azoui

369
2
4

alex · Answer 8 · 2021-06-05T12:32:17.700

0

I have another idea to simple this code if pydantic version below 1.2 that doesn't support parse_obj_as method.

user_list = []
for user in users:
  user_list.append(User(**user))

simple way

user_list = [User(**user) for user in users]

edited Jun 05 '21 at 12:32

answered Jun 05 '21 at 10:47

alex

1
1

How to parse list of models with Pydantic

8 Answers8

Pydantic V1:

Pydantic V2:

Linked