pydantic models¶

If your app already uses pydantic, into= also takes a pydantic v2 model. Install the extra first:

uv add 'fastpg[pydantic]'

import datetime

import pydantic


class Person(pydantic.BaseModel):
    account_id: int = pydantic.Field(alias="id")
    name: str
    created_at: datetime.datetime


query = "SELECT id, name, created_at FROM people WHERE id = $1"
person = await conn.select_one(query, 3, into=Person)
print(person.account_id, person.name)  # 3 person-3

Each column fills the field with the same alias, or the same name if it has no alias. A field with a default can be left out of the query.

Validation¶

fastpg only skips pydantic's validation when it would change nothing. A model with plain field types and no validators or constraints is built directly, which is fast. Any other model is validated by pydantic as usual.

To make sure a model always takes the fast path, pass strategy="direct". It then raises UnsupportedTypeError, with the reason, instead of validating.

A row that fails validation raises fastpg.DecodeError. pydantic's ValidationError is its __cause__.

msgspec Structs are faster than pydantic models, so prefer them for new code.