Class VelocityOptions

Namespace

Datafication.Storage.Velocity

Assembly

Datafication.Storage.Velocity.dll

Configuration options for VelocityDataBlock storage operations

public class VelocityOptions

Inheritance

object

VelocityOptions

Properties

AppendRollThresholdBytes

Max size for a hot appendable segment before rolling. Default: 256MB

public long AppendRollThresholdBytes { get; set; }

Property Value

long

AutoCompactionEnabled

Whether automatic compaction is enabled. Default: false

public bool AutoCompactionEnabled { get; set; }

Property Value

bool

AutoCompactionTrigger

Trigger condition for automatic compaction. Default: SegmentCount

public VelocityCompactionTrigger AutoCompactionTrigger { get; set; }

Property Value

VelocityCompactionTrigger

CompactionBatchSize

Batch size for compaction operations (rows processed per batch). Default: 1m

public int CompactionBatchSize { get; set; }

Property Value

int

CompactionMaxMemoryBytes

Maximum memory usage for compaction operations in bytes. Default: 2GB

public long CompactionMaxMemoryBytes { get; set; }

Property Value

long

CompressionLevel

Compression level (codec-specific). Default: 1 (fast)

public int CompressionLevel { get; set; }

Property Value

int

An integer representing the compression level, where higher values typically provide better compression ratios at the cost of increased CPU usage.

Remarks

The meaning of compression levels varies by codec:

Level	LZ4	Deflate
1	Fast (L00_FAST)	Fastest
2	Balanced (L03_HC)	Optimal
3	High compression (L09_HC)	SmallestSize

CompressionThresholdBytes

Minimum column size threshold for compression (bytes). Default: 1KB

public int CompressionThresholdBytes { get; set; }

Property Value

int

The minimum size in bytes that a column's data must reach before compression is considered beneficial.

Remarks

Columns with data smaller than this threshold will not be compressed, as the overhead of compression may exceed the storage savings for small datasets. This helps optimize performance by avoiding unnecessary compression operations.

DefaultCompression

Default compression type for new columns. Default: None

public VelocityCompressionType DefaultCompression { get; set; }

Property Value

VelocityCompressionType

The compression type to use for columns when auto-compression is disabled. Set to None for no compression, or specify a codec like LZ4 for fast compression.

Remarks

This setting is used when EnableAutoCompression is false, or as a fallback when auto-compression cannot determine a suitable codec.

EnableAutoCompression

Whether to enable automatic compression selection based on data characteristics. Default: false

public bool EnableAutoCompression { get; set; }

Property Value

bool

true to enable automatic compression selection; false to use DefaultCompression for all columns.

Remarks

When enabled, the system analyzes data types and characteristics to select the most appropriate compression codec for each column. For example:

• String columns typically use LZ4 or Deflate compression
• Large fixed-width types (GUID, Decimal) may benefit from compression
• Small types (bool, byte) typically remain uncompressed

EnableAutoMaterialization

Enable automatic WAL materialization based on thresholds. Default: true

public bool EnableAutoMaterialization { get; set; }

Property Value

bool

EnableBackgroundCompaction

Whether to enable background compaction. Default: false

public bool EnableBackgroundCompaction { get; set; }

Property Value

bool

EnableDirectFileCopy

Enable direct file copying for unchanged segments during compaction. Default: true

public bool EnableDirectFileCopy { get; set; }

Property Value

bool

EnableMVCC

Whether to enable MVCC (Multi-Version Concurrency Control). Default: true

public bool EnableMVCC { get; set; }

Property Value

bool

EnableParallelCompaction

Enable parallel processing during compaction when possible. Default: true

public bool EnableParallelCompaction { get; set; }

Property Value

bool

EnableStreamingCompaction

Enable streaming compaction mode for large datasets. Default: true

public bool EnableStreamingCompaction { get; set; }

Property Value

bool

MaxConcurrentCompactionTasks

Maximum number of concurrent compaction tasks. Default: CPU count

public int MaxConcurrentCompactionTasks { get; set; }

Property Value

int

MaxScanDegreeOfParallelism

Optional override for maximum scan degree-of-parallelism during read/filters. Default: null (auto).

public int? MaxScanDegreeOfParallelism { get; set; }

Property Value

int?

MaxSegmentSizeBytes

Maximum size for each segment file in bytes. Default: 48GB

public long MaxSegmentSizeBytes { get; set; }

Property Value

long

MaxSegmentsBeforeCompaction

Maximum number of segments before auto-compaction is triggered. Default: 10

public int MaxSegmentsBeforeCompaction { get; set; }

Property Value

int

PartitioningMode

Preferred partitioning mode for scanning (auto selects best).

public VelocityScanPartitioningMode PartitioningMode { get; set; }

Property Value

VelocityScanPartitioningMode

PrimaryKeyColumn

Primary key column name for CRUD operations. Required for updates/deletes.

public string? PrimaryKeyColumn { get; set; }

Property Value

string

RowGroupTargetBytes

Target size per appended row-group in bytes. Default: 128MB

public long RowGroupTargetBytes { get; set; }

Property Value

long

RowGroupTargetRows

Target rows per appended row-group (when materializing). Default: 64k

public int RowGroupTargetRows { get; set; }

Property Value

int

Strategy

Compaction strategy to use. Default: BinPack

public VelocityCompactionStrategy Strategy { get; set; }

Property Value

VelocityCompactionStrategy

TargetSegmentSizeBytes

Target size for each segment file in bytes. Default: 32GB

public long TargetSegmentSizeBytes { get; set; }

Property Value

long

TombstoneFlushThreshold

How often to flush tombstone records to disk. Default: 1000 records

public int TombstoneFlushThreshold { get; set; }

Property Value

int

WalEnabled

Enable Write-Ahead Log for low-latency single-row updates. Default: true

public bool WalEnabled { get; set; }

Property Value

bool

WalFsyncPolicy

Fsync policy for WAL appends. Default: Batch

public VelocityWalFsyncPolicy WalFsyncPolicy { get; set; }

Property Value

VelocityWalFsyncPolicy

WalMaxAge

Maximum WAL age before materialization is suggested. Default: 10 minutes

public TimeSpan WalMaxAge { get; set; }

Property Value

TimeSpan

WalMaxBytes

Maximum WAL size in bytes before materialization is suggested. Default: 128MB

public long WalMaxBytes { get; set; }

Property Value

long

WalMaxRecords

Maximum WAL record count before materialization is suggested. Default: 100k

public int WalMaxRecords { get; set; }

Property Value

int

Methods

CreateDefault()

Creates default velocity options

public static VelocityOptions CreateDefault()

Returns

VelocityOptions

CreateHighThroughput()

Creates velocity options optimized for high-throughput writes with advanced compaction

public static VelocityOptions CreateHighThroughput()

Returns

VelocityOptions

CreateUpdateOptimized()

Creates velocity options optimized for frequent updates

public static VelocityOptions CreateUpdateOptimized()

Returns

VelocityOptions