Class ParquetBloomFilterOptions

Namespace

Plank.Schema

Assembly

Plank.dll

Configures a split-block Bloom filter for one column chunk.

public sealed record ParquetBloomFilterOptions : IEquatable<ParquetBloomFilterOptions>

Inheritance

object

ParquetBloomFilterOptions

Implements

IEquatable<ParquetBloomFilterOptions>

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.ReferenceEquals(object, object)

object.ToString()

Fields

Default

The Parquet-recommended default Bloom-filter configuration.

public static readonly ParquetBloomFilterOptions Default

Field Value

ParquetBloomFilterOptions

Properties

ExpectedDistinctValueCount

Gets an optional estimate of the number of distinct non-null values in each row group.

public uint? ExpectedDistinctValueCount { get; init; }

Property Value

uint?

Remarks

When omitted, the writer uses the non-null value count. Supplying a good estimate can reduce the filter size for columns containing many duplicate values.

FalsePositiveProbability

Gets the target probability that a missing value is reported as possibly present.

public double FalsePositiveProbability { get; init; }

Property Value

double

MaximumBytes

Gets the maximum number of bytes used by a column chunk's Bloom-filter bitset.

public uint MaximumBytes { get; init; }

Property Value

uint

Remarks

The value must be a power of two between 32 bytes and 128 MiB, inclusive.