TPC Patterns Usage Guide¶
benchbox.core.tpc_patterns used to ship a generic concurrent-stream execution
framework (StreamExecutor, QueryPermutator, ParameterManager,
TransactionManager, ErrorHandler, ProgressTracker, ResultAggregator,
BenchmarkTestRunner, and related dataclasses/utilities). It was never wired
into the TPC-H or TPC-DS execution paths – the only external consumers of the
module imported generate_official_benchmark_audit_trail, and the framework
duplicated (without ever running) the same stream-execution concept that
benchbox.core.throughput.runner.StreamRunner implements for real. The
framework has been removed; tpc_patterns now contains only
generate_official_benchmark_audit_trail.
What to use instead¶
Power Test:
TPCHBenchmark.run_power_test(...)/TPCDSBenchmark.run_power_test(...)execute all queries sequentially against a real connection and compute Power@Size.Throughput Test:
benchbox.core.tpch.throughput_test.TPCHThroughputTestandbenchbox.core.tpcds.throughput_test.TPCDSThroughputTestexecute multiple concurrent query streams for real, via the sharedbenchbox.core.throughput.runner.StreamRunnerexecutor, and compute Throughput@Size.Official benchmark audit trail:
generate_official_benchmark_audit_trailinbenchbox.core.tpc_patternsis unchanged and still used bybenchbox.core.tpch.official_benchmarkandbenchbox.core.tpcds.official_benchmarkto write the shared audit-trail report format.
from benchbox.core.tpc_patterns import generate_official_benchmark_audit_trail
generate_official_benchmark_audit_trail(
result=official_result,
benchmark_title="TPC-H",
benchmark_slug="tpch",
qph_label="QphH@Size",
qph_attr="qphh_at_size",
)