planframe-pandas 1.3.0

planframe-pandas

pandas adapter package for PlanFrame. Import as planframe_pandas.

Documentation (ReadTheDocs):

• pandas track (end users): https://planframe.readthedocs.io/en/latest/planframe_pandas/
• Light API reference: https://planframe.readthedocs.io/en/latest/planframe_pandas/reference/api/

Install

pip install planframe-pandas

Usage

from planframe_pandas import PandasFrame


class User(PandasFrame):
    id: int
    age: int


pf = User([{"id": 1, "age": 10}, {"id": 2, "age": 20}])
rows = pf.select("id").to_dicts()

Notes

• Core v1.2.0+ (planframe, current minor v1.3.x): shared helpers include execute_plan_async, planframe.materialize, Expr operator IR, Frame *_async aliases, and v1.3.0 adapter/typing notes—see Migrating since v1.1.0.
• PandasFrame subclasses core planframe.pandas.PandasLikeFrame, so you get pandas-flavored helpers (assign, sort_values, boolean indexing, column filter, astype, eval, drop_duplicates, …) on top of the shared Frame plan. See the pandas-like API guide.
• PlanFrame is always lazy: chaining does not touch backend data; execution happens at collect() boundaries.
• collect() returns list[pydantic.BaseModel]. Use collect_backend() for a pandas.DataFrame, or stream_dicts() / stream() to iterate rows (see Streaming rows).

Common transforms

• with_row_index(name="row_nr", offset=0): add a monotonically increasing row number column.
• clip(lower=..., upper=..., subset=...): clamp numeric columns (if subset=None, clamps all numeric schema fields).
• cast_many / cast_subset: multi-column cast helpers.
• fill_null_many / fill_null_subset: multi-column fill-null helpers.
• rename_upper/lower/title/strip(...): schema-driven rename helpers.
• pivot_longer(...) / pivot_wider(...): reshape convenience wrappers around melt / pivot.