`read_json` silently drops offset, unlike `read_csv`

Author: MarcoGorelliCreated Sep 11, 2026Updated Sep 16, 2026

LabelspythonA-temporalacceptedP-highA-io-json

Checks

I have checked that this issue has not already been reported.
I have confirmed this bug exists on the latest version of Polars.

Reproducible example

python

import polars as pl
import io

# I'd expect the following two to give different results...
print(pl.read_json(io.StringIO('[{"a": "2020-01-01T12:34+05:00"}]'), schema={'a': pl.Datetime}))
print(pl.read_json(io.StringIO('[{"a": "2020-01-01T12:34+06:00"}]'), schema={'a': pl.Datetime}))

Log output

bash

# but no, they give the same exact output
shape: (1, 1)
┌─────────────────────┐
│ a                   │
│ ---                 │
│ datetime[μs]        │
╞═════════════════════╡
│ 2020-01-01 12:34:00 │
└─────────────────────┘
shape: (1, 1)
┌─────────────────────┐
│ a                   │
│ ---                 │
│ datetime[μs]        │
╞═════════════════════╡
│ 2020-01-01 12:34:00 │
└─────────────────────┘

Issue description

Different inputs shouldn't result in the same output, it's quite odd that the offset is just silently dropped

Shouldn't the result be converted to UTC, like how read_csv does?

python

>>> pl.read_csv(io.StringIO('a\n2020-01-01T12:34+05:00\n2020-01-01T12:34+06:00'), try_parse_dates=True)
Out[16]:
shape: (2, 1)
┌─────────────────────────┐
│ a                       │
│ ---                     │
│ datetime[μs, UTC]       │
╞═════════════════════════╡
│ 2020-01-01 07:34:00 UTC │
│ 2020-01-01 06:34:00 UTC │
└─────────────────────────┘

Expected behavior

I think

shape: (1, 1)
┌─────────────────────┐
│ a                   │
│ ---                 │
│ datetime[μs]        │
╞═════════════════════╡
│ 2020-01-01 07:34:00 │
└─────────────────────┘
shape: (1, 1)
┌─────────────────────┐
│ a                   │
│ ---                 │
│ datetime[μs]        │
╞═════════════════════╡
│ 2020-01-01 06:34:00 │
└─────────────────────┘

Installed versions

--------Version info---------
Polars:              2.0.0-rc.1
Index type:          UInt32
Platform:            Linux-6.6.87.2-microsoft-standard-WSL2-x86_64-with-glibc2.39
Python:              3.14.4 (main, Apr 14 2026, 14:26:14) [Clang 22.1.3 ]
Runtime:             rt32

----Optional dependencies----
Azure CLI            <not installed>
adbc_driver_manager  <not installed>
altair               <not installed>
azure.identity       <not installed>
boto3                <not installed>
cloudpickle          <not installed>
connectorx           <not installed>
deltalake            <not installed>
fastexcel            <not installed>
fsspec               <not installed>
gevent               <not installed>
google.auth          <not installed>
great_tables         <not installed>
matplotlib           <not installed>
numpy                2.4.6
openpyxl             <not installed>
pandas               3.0.3
polars_cloud         <not installed>
pyarrow              24.0.0
pydantic             2.13.5
pyiceberg            <not installed>
sqlalchemy           <not installed>
torch                <not installed>
xlsx2csv             <not installed>
xlsxwriter           <not installed>

Source: pola-rs/polars

View original on GitHub View discussion on GitHub