# public_bi_benchmark **Repository Path**: mirrors_pingcap/public_bi_benchmark ## Basic Information - **Project Name**: public_bi_benchmark - **Description**: BI benchmark with user generated data and queries - **Primary Language**: Unknown - **License**: MIT - **Default Branch**: master - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2020-08-19 - **Last Updated**: 2025-10-19 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # Public BI benchmark for TiDB ## Data Size There are 46 workbooks containing 206 tables (.csv files) with the total size of 41 GB compressed and 386 GB uncompressed. ## Usage ``` go run main.go [download|prepare|run|cleanup]: //options used to connect the database server. -P int MySQL Port (default 4000) -d string MySQL database (default "test") -h string MySQL Host (default "127.0.0.1") -u string MySQL User (default "root") -p string MySQL Password //options used to control the test behavior. -c value test case you want to benchmark -i string input benchmark resource (default "./benchmark") -o string output directory to save cvs files and logs (default "./var") -sample Use sample to test -ignore-run-error Ignore errors when running ``` Hint: 1. When using sample to test, you can run `prepare` directly without `download` . 2. When using datasets to test, download datasets in other ways first, use options `-o` to select file directory. (Maybe the download speed is too slow, VPN or downloading from internal server can work well.) ## Status For running the benchmark in TiDB, some datasets and SQL queries are adjusted. The adjustment is as follows. ### Deleted datasets: Datasets 'Wins' and 'USCensus' (The number of columns is greater than 512.)are deleted. * In workbook 'Wins' , there are 13 queries. * In workbook 'USCensus' , there are 8 queries. ### Adjusted SQL queries: Change `cast xx as bigint` to `cast xx as signed` ### Deleted SQL queries: N*interval: ``` benchmark/CommonGovernment/queries/2.sql benchmark/CommonGovernment/queries/22.sql benchmark/CommonGovernment/queries/23.sql ``` trim,splitpart: ``` benchmark/IGlocations1/queries/16.sql benchmark/Romance/queries/8.sql (only trim) benchmark/Romance/queries/9.sql (only trim) ``` median: ``` benchmark/PanCreactomy1/queries/2.sql benchmark/PanCreactomy2/queries/4.sql ``` cast as timestamp: ``` benchmark/RealEstate2/queries/31.sql ``` cast as text: ``` benchmark/TrainsUK1/queries/6.sql benchmark/TrainsUK1/queries/7.sql benchmark/Uberlandia/queries/12.sql ```