KO
|
EN
gitlite — search
Search
#python
#golang
#python3
#hacktoberfest
#javascript
#docker
#cli
#bot
#github-actions
#android
#images
#reactjs
Testing_BI_Engine
★ 53
Open GitHub ↗
TPC-H_SF10
Download README (.md)
Explore Similar Repositories
alloydb
:
AlloyDB is a distributed SQL database.
alloydb-go-connector
:
A Go library for connecting securely to your AlloyDB instances
pytest-databases
:
Reusable database fixtures for any and all databases.
alloydb-python-connector
:
A Python library for connecting securely to your AlloyDB instances.
alloydb
:
Skills for AlloyDB
// repository documentation
Was this content helpful?
★ 0
(0 ratings)
Select Rating:
★
★
★
★
★
Submit Feedback
Recent Feedback
×
Download README
Do you want to download the
README.md
file for
Testing_BI_Engine
?
Download (.md)
# Update : added results for SF100 Blog : https://datamonkeysite.com/2023/03/09/benchmarking-snowflake-databricks-synapse-bigquery-and-duckdb-using-tpch-sf100/ # Testing OLAP SQL Engine using TPC-H it started as testing BigQuery BI engine for fun, but I did found it easy to test other Engine too, Please check this blog for Background https://datamonkeysite.com/2022/01/07/benchmark-snowflake-bigquery-singlestore-and-databricks-using-tpc-h-sf10/ so Far Tested Snowflake, BigQuery, Databrciks SQL, SingleStore,DuckDB, and PowerBI Datamart, I only include DB that take less than 100 second to finish the benchmark it is not supposed to be very accurate or an official benchmark, but just to get a general impression of the performance. TPC-H reference : http://tpc.org/tpc_documents_current_versions/pdf/tpc-h_v3.0.0.pdf # BigQuery BI Engine The Original Purpose of the Benchmark was to troll BiQuery Product team to make all Queries accelerated by BI Engine :) Unfortunately not all Queries are accelerated by BigQuery BI Engine (Currently only 11) Currently BI Engine is limited to a join with 5 Million Dimension Table up to 5 Joins are supported see Current Limitations https://cloud.google.com/bi-engine/docs/sql-interface-overview#limitations TPC-H SF10 use up to 8 joins and a dimension Table with 15 Million Records. Results for BI Engine Only : https://datastudio.google.com/reporting/b162bcc1-baee-4af5-995f-6155c939d742/page/p_v72vuqfvqc # PowerBI Datamart it is the easiest to setup, loading data uses only GUI, I love it, but it is a read only DB, you can can't delete rows etc #Apache Spark added Pyspark, just for fun, all I can say, it is really slow in a single node system
