ldbc_data_gen

Generate LDBC SNB BI datasets directly from DuckDB

Maintainer(s): Dtenwolde

Installing and Loading

INSTALL ldbc_data_gen FROM community;
LOAD ldbc_data_gen;

Example

INSTALL ldbc_data_gen FROM community;
LOAD ldbc_data_gen;

-- Generate the LDBC SNB BI dataset as DuckDB tables.
CALL ldbcgen(sf := 0.003, schema := 'ldbc');

-- Inspect generated relation counts.
FROM ldbcgen(sf := 0.003, schema := 'ldbc_counts');

-- Generate Spark-compatible BI Parquet files.
CALL ldbcgen(
  sf := 0.003,
  target := 'files',
  output_dir := 'out/sf0.003',
  format := 'parquet',
  overwrite := true
);

-- Inspect the generated schema metadata.
FROM ldbcgen_schema(format := 'parquet')
LIMIT 10;

-- Run one bundled LDBC BI query against generated tables.
PRAGMA ldbc_bi(1, schema = 'ldbc');

About ldbc_data_gen

The ldbc_data_gen extension generates deterministic LDBC Social Network Benchmark Business Intelligence (SNB BI) datasets directly inside DuckDB. It can materialize the generated graph as DuckDB tables or write Spark-compatible BI files in the composite-merged-fk layout, including the initial snapshot, inserts, and deletes. Parquet and CSV file output are supported. The extension also includes helpers for inspecting the generated schema, configuration values, deterministic Java-compatible random sequences, and bundled LDBC BI query execution through PRAGMA ldbc_bi. The generator is designed for reproducible benchmark data generation and local validation workflows without requiring Spark.

Added Functions

Overloaded Functions

This extension does not add any function overloads.

Added Types

This extension does not add any types.

Added Settings

This extension does not add any settings.