Getting Started¶

Installation¶

pip install polars-genson[polars]

import polars as pl
import polars_genson

df = pl.DataFrame({
    "json_data": [
        '{"name": "Alice", "age": 30, "scores": [95, 87]}',
        '{"name": "Bob", "age": 25, "city": "NYC", "active": true}'
    ]
})

json_schema = df.genson.infer_json_schema("json_data")
polars_schema = df.genson.infer_polars_schema("json_data")

Install (dev)¶

cd polars-genson-py/
uv venv && source .venv/bin/activate
uv sync

Troubleshooting¶

Memory usage exceeds RAM¶

If memory usage exceeds RAM, try reducing the max. schema builders.

For example, if you have a DataFrame with 1000 rows, and call .genson.normalise_json on it, you'll by default get 1000 threads that get scheduled on your available cores. This will mean that at one moment in time you will have 1000 genson-rs schema "builders" all storing their built schemas, before they are all merged in one go. If you limit to 100 builders, they will be merged 100 at a time, which will reduce the peak RSS (RAM use by the process).