DataFrameSchema

laktory.models.dataframe.DataFrameSchema

Bases: BaseModel

DataFrame schema. Typically used to explicitly express a schema when reading files.

Examples:

from laktory import models

schema = models.DataFrameSchema(columns={"a": "string", "x": "double"})

PARAMETER	DESCRIPTION
columns	Dict or list of columns TYPE: dict[str, str | DType | DataFrameColumn] | list[DataFrameColumn] | VariableType

METHOD	DESCRIPTION
from_df	Create a DataFrameSchema from a DataFrame
from_narwhals	Create a DataFrameSchema from a Narwhals schema
to_narwhals	Returns a Narwhals schema object
to_polars	Returns a Polars schema object
to_pyspark	Returns a Spark schema object
to_string	Returns a string representation of the schema

from_df(df) classmethod

Create a DataFrameSchema from a DataFrame

Columns are populated from the narwhals schema for introspection. The native schema is stored directly so that to_native() can return it as-is, bypassing column-by-column conversion.

Source code in laktory/models/dataframe/dataframeschema.py

@classmethod
def from_df(cls, df) -> "DataFrameSchema":
    """
    Create a DataFrameSchema from a DataFrame

    Columns are populated from the narwhals schema for introspection. The
    native schema is stored directly so that `to_native()` can return it as-is,
    bypassing column-by-column conversion.
    """
    if not isinstance(df, (nw.LazyFrame, nw.DataFrame)):
        df = df.from_native(df)

    obj = cls.from_narwhals(df.collect_schema())
    obj._native_schema = df.to_native().schema

    return obj

from_narwhals(schema, native_schema=None) classmethod

Create a DataFrameSchema from a Narwhals schema

Source code in laktory/models/dataframe/dataframeschema.py

@classmethod
def from_narwhals(cls, schema: nw.Schema, native_schema=None) -> "DataFrameSchema":
    """Create a DataFrameSchema from a Narwhals schema"""
    columns = [
        DataFrameColumn(name=name, dtype=DType.from_narwhals(dtype))
        for name, dtype in schema.items()
    ]
    obj = cls(columns=columns)
    if native_schema:
        obj._native_schema = native_schema

    return obj

to_narwhals()

Returns a Narwhals schema object

Source code in laktory/models/dataframe/dataframeschema.py

def to_narwhals(self) -> nw.Schema:
    """Returns a Narwhals schema object"""
    cols = {}
    for c in self.columns:
        cols[c.name] = c.dtype.to_narwhals()
    return nw.Schema(cols)

to_polars()

Returns a Polars schema object

Source code in laktory/models/dataframe/dataframeschema.py

def to_polars(self):
    """Returns a Polars schema object"""
    import polars as pl

    cols = {}
    for c in self.columns:
        cols[c.name] = c.dtype.to_polars()
    return pl.Schema(cols)

to_pyspark()

Returns a Spark schema object

Source code in laktory/models/dataframe/dataframeschema.py

def to_pyspark(self):
    """Returns a Spark schema object"""
    import pyspark.sql.types as T

    columns = []
    for c in self.columns:
        _type = c.dtype.to_pyspark()
        columns += [T.StructField(c.name, _type, c.nullable)]

    return T.StructType(columns)

to_string(indent=None)

Returns a string representation of the schema

Source code in laktory/models/dataframe/dataframeschema.py

def to_string(self, indent=None):
    """Returns a string representation of the schema"""
    return json.dumps(
        {c.name: c.dtype.to_string() for c in self.columns}, indent=indent
    )