API Reference¶

Overview of the pyfloe public API, organized by topic.

Quick reference¶

Function	Description
`LazyFrame(data)`	From list of dicts or objects
`read_csv(path, ...)`	Lazy CSV reader (auto-detects datetime)
`read_tsv(path, ...)`	Lazy TSV reader
`read_jsonl(path, ...)`	Lazy JSON Lines reader
`read_json(path, ...)`	JSON array reader
`read_fixed_width(path, widths, ...)`	Lazy fixed-width reader
`read_parquet(path, ...)`	Lazy Parquet reader (requires pyarrow)
`from_iter(source, ...)`	From any iterator/generator
`from_chunks(chunks, ...)`	From batched/paginated source
`Stream.from_iter(source, ...)`	True streaming pipeline
`Stream.from_csv(path, ...)`	Stream from CSV

Method	Lazy?	Description
`.select(*cols)`	✓	Select columns or expressions
`.filter(expr)`	✓	Filter rows
`.with_column(name, expr)`	✓	Add computed column
`.with_columns(**exprs)`	✓	Add multiple columns
`.drop(*cols)`	✓	Drop columns
`.rename(mapping)`	✓	Rename columns
`.sort(*cols)`	✗	Sort (Timsort)
`.join(other, on=)`	✓	Hash join (or sort-merge with `sorted=True`)
`.union(other)`	✓	Stack rows
`.explode(col)`	✓	Unnest lists
`.apply(func)`	✓	Apply to columns
`.group_by(*cols).agg(...)`	✗	Hash agg (or streaming with `sorted=True`)
`.head(n)`	partial	First n rows
`.optimize()`	✓	Optimized plan
`.collect()`	✗	Materialize
`.to_pylist()`	✗	→ List[dict]
`.to_csv(path)`	streaming	Write CSV
`.to_jsonl(path)`	streaming	Write JSONL
`.explain()`	✓	Print plan
`.schema`	✓	Schema (no data)
`.typed(T)`	✓	→ TypedLazyFrame[T]
`.validate(T)`	✓	Check schema

Accessor	Methods
`.str`	`upper`, `lower`, `strip`, `title`, `len`, `contains`, `startswith`, `endswith`, `replace`, `slice`
`.dt`	`year`, `month`, `day`, `hour`, `minute`, `second`, `microsecond`, `weekday`, `isoweekday`, `quarter`, `week`, `day_of_year`, `day_name`, `month_name`, `date`, `time`, `truncate`, `strftime`, `epoch_seconds`, `add_days`, `add_hours`, `add_minutes`, `add_seconds`