Hybrid Columnar Compression (HCC) User Guide

Hybrid Columnar Compression (HCC) User Guide

Overview

HCC provides Oracle Exadata-style compression with configurable algorithms (ZSTD, LZ4, Snappy) achieving 3-10x compression ratios.

Benefits

• 3-10x compression ratio
• Reduced storage costs
• Faster scans (less I/O)
• Configurable per table/column

Configuration

-- Table-level compression
CREATE TABLE events (
    timestamp TIMESTAMPTZ,
    data JSONB
) WITH (compression = 'zstd', compression_level = 3);

-- Column-specific compression
CREATE TABLE mixed (
    id INT,
    name TEXT COMPRESSION lz4,
    data BYTEA COMPRESSION zstd
);

Algorithms

Algorithm	Ratio	Speed	Use Case
ZSTD	10x	Medium	Archives
LZ4	3x	Fast	OLTP
Snappy	2x	Fastest	Real-time

SQL Examples

-- Check compression ratio
SELECT
    tablename,
    uncompressed_size_mb,
    compressed_size_mb,
    ROUND(uncompressed_size_mb / compressed_size_mb, 2) as ratio
FROM heliosdb.compression_stats;

-- Change compression
ALTER TABLE my_table SET compression = 'zstd';
VACUUM FULL my_table;  -- Apply compression

Best Practices

1. Use ZSTD for cold data
2. Use LZ4 for hot data
3. Monitor compression ratios
4. Recompress periodically

For more: /docs/performance/compression.md