API

Import each name from astrolabe.

Frames and schemas

Name	Meaning
TypedDataFrame[Schema, Rest]	A pl.DataFrame whose type states its columns. Rest is optional.
TypedLazyFrame[Schema, Rest]	A pl.LazyFrame whose type states its columns.
TypedSeries[Name, DType]	A pl.Series whose type states its name and its data type.
Columns	The base class of a schema. Each field is a column.
columns(name, fields)	Declare a schema from a dict, for column names that are not Python identifiers.
Column[Name, DType]	One column, with its name and its data type, in a type. A field of Struct[...] uses it.

See Schemas.

Data types

Datetime, Duration, Decimal, Enum, List, Array and Struct name a data type with parameters in a type annotation. Unchecked, Planned, AnyZone and EveryDType are special data types. NumericType, IntegerType, SignedIntegerType, UnsignedIntegerType, FloatType, TemporalType and NestedType are the families of Polars. See Data types.

Remainders

Name	Meaning
other_columns(...)	Declare a remainder, with its promises.

For the remainder of a result, see InferredRemainder below, and Remainders.

Declarations

Name	Meaning
column(name, dtype)	An expression that reads one column and gives it back.
typed_expr(*reads, gives=, named=)	An expression that gives one value for each row.
typed_agg(*reads, gives=, named=)	An expression that gives one value for each group.
typed_lit(gives=, named=)	An expression that gives one fixed value and reads no column.
typed_selection(selector, gives=, suffix=, prefix=)	A selection that gives one value for each row.
typed_agg_selection(selector, gives=)	A selection that gives one value for each group.
row(*reads, gives=)	One result for some input data types.
some_column(region)	A read of one column whose name each call chooses.
same	The result: the data type of the column that the expression reads.
wrap(constructor, inner)	The result: a nested data type around another result.
column_of(Schema, name), subset_of(Schema, *names)	Reads of the columns of a Columns class.
TypedExpr, TypedColumn, TypedAgg, TypedLit	The types that the declarations and the checker give to expressions.

See Declare expression types.

Inferred return types

Name	Meaning
InferredExpr, InferredDataFrame, InferredLazyFrame	A return type that the checker computes from the body.
InferredRemainder	A remainder that the checker computes from the body.
infer(value)	Mark one returned value.

See Inferred return types.

Runtime functions

Name	At run time
check_schema(df, Frame)	Check that the frame has exactly the declared columns, and type it.
force_schema(df, Frame, *, strict=True)	Keep the declared columns, cast each one, and type the frame.
assume_schema(df, Frame)	Type the frame, with no check.

See Runtime checks.

Markers in revealed types

The checker writes a remainder that it computes with these names, in a revealed type and in a message. A program does not need to import them: write InferredRemainder and infer(...).

Name	Meaning
Minus, Only, Except, Suffixed, Prefixed, Retyped, Both, SplitColumns	The markers of a remainder.
OneOf, StartsWith, EndsWith, Contains, Matching, OfType, SelAnd, SelOr, SelNot	The markers of a selector.

See Markers in revealed types.