ExDataSketch.Theta (ExDataSketch v0.7.1)

Copy Markdown View Source

Theta Sketch for set operations on cardinalities.

Theta sketches support cardinality estimation with set operations (union, intersection, difference) that other sketch families do not natively support. This makes them ideal for queries like "how many users visited both page A and page B?"

How It Works

A Theta sketch maintains a set of hash values below a threshold (theta). When the set exceeds the nominal size k, the threshold is lowered and entries above it are discarded. The cardinality estimate is derived from the number of retained entries and the current theta value.

Options

  • :k - nominal number of entries (default: 4096). Controls accuracy. Higher values use more memory but give better estimates. Must be a power of 2, between 16 and 67,108,864 (2^26).
  • :backend - backend module (default: ExDataSketch.Backend.Pure).

Binary State Layout (v1)

All multi-byte fields are little-endian.

Offset  Size    Field
------  ------  -----
0       1       Version (u8, currently 1)
1       4       k nominal entries (u32 little-endian)
5       8       Theta value (u64 little-endian, max = 2^64-1 = "no threshold")
13      4       Entry count (u32 little-endian)
17      N*8     Entries (sorted array of u64 little-endian hash values)

Total: 17 + entry_count * 8 bytes.

DataSketches Interop

Theta is the primary target for Apache DataSketches interop. serialize_datasketches/1 and deserialize_datasketches/1 implement the CompactSketch binary format, enabling cross-language compatibility with Java, C++, and Python DataSketches libraries.

Merge Properties

Theta merge (union) is associative and commutative. This means sketches can be merged in any order or grouping and produce the same result, making Theta safe for parallel and distributed aggregation.

Summary

Functions

Compacts the sketch into a read-only form with sorted entries.

Deserializes an EXSK binary into a Theta sketch.

Deserializes an Apache DataSketches CompactSketch binary into a Theta sketch.

Estimates the cardinality (distinct count) from the sketch.

Creates a new Theta sketch from an enumerable of items.

Merges two Theta sketches (set union).

Merges a non-empty enumerable of Theta sketches into one.

Returns a 2-arity merge function suitable for combining sketches.

Creates a new Theta sketch.

Returns a 2-arity reducer function suitable for Enum.reduce/3 and similar.

Serializes the sketch to the ExDataSketch-native EXSK binary format.

Serializes the sketch to Apache DataSketches CompactSketch format.

Returns the size of the sketch state in bytes.

Updates the sketch with a single item.

Updates the sketch with multiple items in a single pass.

Types

t()

@type t() :: %ExDataSketch.Theta{backend: module(), opts: keyword(), state: binary()}

Functions

compact(sketch)

@spec compact(t()) :: t()

Compacts the sketch into a read-only form with sorted entries.

Compacting discards any entries above the current theta threshold and sorts the remaining entries. This is required before serialization to the DataSketches CompactSketch format.

Examples

iex> sketch = ExDataSketch.Theta.new() |> ExDataSketch.Theta.update("x") |> ExDataSketch.Theta.compact()
iex> ExDataSketch.Theta.estimate(sketch) > 0.0
true

deserialize(binary)

@spec deserialize(binary()) :: {:ok, t()} | {:error, Exception.t()}

Deserializes an EXSK binary into a Theta sketch.

Returns {:ok, sketch} on success or {:error, reason} on failure.

Examples

iex> ExDataSketch.Theta.deserialize(<<"invalid">>)
{:error, %ExDataSketch.Errors.DeserializationError{message: "deserialization failed: invalid magic bytes, expected EXSK"}}

deserialize_datasketches(binary, opts \\ [])

@spec deserialize_datasketches(
  binary(),
  keyword()
) :: {:ok, t()} | {:error, Exception.t()}

Deserializes an Apache DataSketches CompactSketch binary into a Theta sketch.

Options

  • :seed - expected seed value for seed hash verification (default: 9001).

Examples

iex> sketch = ExDataSketch.Theta.new(k: 1024) |> ExDataSketch.Theta.update("test")
iex> binary = ExDataSketch.Theta.serialize_datasketches(sketch)
iex> {:ok, restored} = ExDataSketch.Theta.deserialize_datasketches(binary)
iex> ExDataSketch.Theta.estimate(restored) == ExDataSketch.Theta.estimate(sketch)
true

estimate(theta)

@spec estimate(t()) :: float()

Estimates the cardinality (distinct count) from the sketch.

Examples

iex> ExDataSketch.Theta.new() |> ExDataSketch.Theta.estimate()
0.0

from_enumerable(enumerable, opts \\ [])

@spec from_enumerable(
  Enumerable.t(),
  keyword()
) :: t()

Creates a new Theta sketch from an enumerable of items.

Equivalent to new(opts) |> update_many(enumerable).

Options

Same as new/1.

Examples

iex> sketch = ExDataSketch.Theta.from_enumerable(["a", "b", "c"], k: 1024)
iex> ExDataSketch.Theta.estimate(sketch) > 0.0
true

merge(sketch, theta)

@spec merge(t(), t()) :: t()

Merges two Theta sketches (set union).

Both sketches must have the same k value. Returns the merged sketch. Raises ExDataSketch.Errors.IncompatibleSketchesError if the sketches have different parameters.

Examples

iex> a = ExDataSketch.Theta.new(k: 1024) |> ExDataSketch.Theta.update("x")
iex> b = ExDataSketch.Theta.new(k: 1024) |> ExDataSketch.Theta.update("y")
iex> merged = ExDataSketch.Theta.merge(a, b)
iex> ExDataSketch.Theta.estimate(merged) >= ExDataSketch.Theta.estimate(a)
true

merge_many(sketches)

@spec merge_many(Enumerable.t()) :: t()

Merges a non-empty enumerable of Theta sketches into one.

Raises Enum.EmptyError if the enumerable is empty.

Examples

iex> a = ExDataSketch.Theta.new(k: 1024) |> ExDataSketch.Theta.update("x")
iex> b = ExDataSketch.Theta.new(k: 1024) |> ExDataSketch.Theta.update("y")
iex> merged = ExDataSketch.Theta.merge_many([a, b])
iex> ExDataSketch.Theta.estimate(merged) > 0.0
true

merger(opts \\ [])

@spec merger(keyword()) :: (t(), t() -> t())

Returns a 2-arity merge function suitable for combining sketches.

The returned function calls merge/2 on two sketches.

Examples

iex> is_function(ExDataSketch.Theta.merger(), 2)
true

new(opts \\ [])

@spec new(keyword()) :: t()

Creates a new Theta sketch.

Options

  • :k - nominal number of entries (default: 4096). Must be a power of 2, between 16 and 67108864.
  • :backend - backend module (default: ExDataSketch.Backend.Pure).
  • :hash_fn - custom hash function (term -> non_neg_integer).
  • :seed - hash seed (default: 0).

Examples

iex> sketch = ExDataSketch.Theta.new(k: 1024)
iex> sketch.opts[:k]
1024
iex> ExDataSketch.Theta.size_bytes(sketch)
17

reducer()

@spec reducer() :: (term(), t() -> t())

Returns a 2-arity reducer function suitable for Enum.reduce/3 and similar.

The returned function calls update/2 on each item.

Examples

iex> is_function(ExDataSketch.Theta.reducer(), 2)
true

serialize(theta)

@spec serialize(t()) :: binary()

Serializes the sketch to the ExDataSketch-native EXSK binary format.

The serialized binary includes magic bytes, version, sketch type, parameters, and state. See ExDataSketch.Codec for format details.

Examples

iex> sketch = ExDataSketch.Theta.new(k: 1024)
iex> binary = ExDataSketch.Theta.serialize(sketch)
iex> <<"EXSK", _rest::binary>> = binary
iex> byte_size(binary) > 0
true

serialize_datasketches(sketch, opts \\ [])

@spec serialize_datasketches(
  t(),
  keyword()
) :: binary()

Serializes the sketch to Apache DataSketches CompactSketch format.

This is the primary interop target for cross-language compatibility. The CompactSketch format uses 64-bit hashes with a seed hash for compatibility verification.

Options

  • :seed - the seed value for seed hash computation (default: 9001).

Examples

iex> sketch = ExDataSketch.Theta.new(k: 1024) |> ExDataSketch.Theta.update("hello")
iex> binary = ExDataSketch.Theta.serialize_datasketches(sketch)
iex> is_binary(binary) and byte_size(binary) > 0
true

size_bytes(theta)

@spec size_bytes(t()) :: non_neg_integer()

Returns the size of the sketch state in bytes.

Examples

iex> ExDataSketch.Theta.new() |> ExDataSketch.Theta.size_bytes()
17

update(sketch, item)

@spec update(t(), term()) :: t()

Updates the sketch with a single item.

The item is hashed using ExDataSketch.Hash.hash64/1 before being inserted into the sketch.

Examples

iex> sketch = ExDataSketch.Theta.new() |> ExDataSketch.Theta.update("hello")
iex> ExDataSketch.Theta.estimate(sketch) > 0.0
true

update_many(sketch, items)

@spec update_many(t(), Enumerable.t()) :: t()

Updates the sketch with multiple items in a single pass.

More efficient than calling update/2 repeatedly because it minimizes intermediate binary allocations.

Examples

iex> sketch = ExDataSketch.Theta.new() |> ExDataSketch.Theta.update_many(["a", "b", "c"])
iex> ExDataSketch.Theta.estimate(sketch) > 0.0
true