Parquex.Dataset.Stream (parquex v0.3.0)

View Source

A pull-based, prefix-pruned stream across dataset parts.

Summary

Functions

Closes the active file reader. This operation is idempotent.

Returns the projected output schema without listing or opening objects.

Returns bounded-cardinality planning and traversal counts without object keys.

Types

t()

@opaque t()

Functions

close(stream)

@spec close(t()) :: :ok | {:error, Parquex.Error.t()}

Closes the active file reader. This operation is idempotent.

open(dataset, options)

@spec open(
  Parquex.Dataset.t(),
  keyword()
) :: {:ok, t()} | {:error, Parquex.Error.t()}

schema(stream)

@spec schema(t()) :: Parquex.Schema.t()

Returns the projected output schema without listing or opening objects.

stats(stream)

@spec stats(t()) :: {:ok, map()}

Returns bounded-cardinality planning and traversal counts without object keys.