column_index.hpp File Reference

ColumnIndex, OffsetIndex, and ColumnIndexBuilder for predicate pushdown. More...

#include "signet/thrift/compact.hpp"
#include "signet/statistics.hpp"
#include <cstdint>
#include <cstring>
#include <string>
#include <vector>

Go to the source code of this file.

Classes
struct	signet::forge::PageLocation
	File offset and size descriptor for a single data page. More...
struct	signet::forge::OffsetIndex
	Page locations for random access within a column chunk. More...
struct	signet::forge::ColumnIndex
	Per-page min/max statistics for predicate pushdown. More...
class	signet::forge::ColumnIndexBuilder
	Builder that accumulates per-page statistics during column writing. More...

Namespaces
namespace	signet
namespace	signet::forge

Detailed Description

ColumnIndex, OffsetIndex, and ColumnIndexBuilder for predicate pushdown.

Per-column-chunk structures for predicate pushdown and random page access. Written after the row groups but before the footer in the Parquet file.

ColumnIndex stores per-page min/max statistics, enabling readers to skip pages that cannot contain matching data. OffsetIndex stores page locations for efficient random access into column chunks.

Thrift field IDs follow the canonical parquet.thrift specification:

• ColumnIndex: 1=null_pages, 2=min_values, 3=max_values, 4=boundary_order, 5=null_counts
• OffsetIndex: 1=page_locations (list of PageLocation)
• PageLocation: 1=offset, 2=compressed_page_size, 3=first_row_index

Definition in file column_index.hpp.