How apache/parquet-format works
This repository defines the Parquet file format and its Thrift metadata structures; it is a specification, not a reader/writer implementation. The documented workflow is that writers place column data before file metadata, while readers inspect metadata to locate and sequentially read selected column chunks. The graph includes documented format extensions and representations. Source excerpts were unavailable, so relationships are limited to those stated in the README.
Main entry points: README.md File layout · parquet.thrift Thrift metadata · LogicalTypes.md Logical types
Components 18
File Structure
- File layout
README.md - Thrift metadata
src/main/thrift/parquet.thrift - Row groups and pages
README.md
Data Representation
- Primitive types
README.md - Logical types
LogicalTypes.md - Nested encoding
README.md
Format Capabilities
- Encodings
Encodings.md - Compression
Compression.md - Encryption
Encryption.md - Bloom filter
BloomFilter.md - Page index
PageIndex.md - Variant encoding
VariantEncoding.md - Variant shredding
VariantShredding.md - Geospatial types
Geospatial.md - ALP encoding
AlpEncoding.md - Binary extensions
BinaryProtocolExtensions.md
Other components
- Format writers
- Format readers
Connections 7
- Format writers → File layout: write format
README.md - File layout → Thrift metadata: defines metadata
README.md - File layout → Row groups and pages: organizes data
README.md - Format readers → Thrift metadata: read first
README.md - Format readers → File layout: locate chunks
README.md - Logical types → Primitive types: annotate types
README.md - Nested encoding → Row groups and pages: encode nesting
README.md