- Suhas Chikkanaravangala Vijayakumar
- Ravikiran Jois Yedur Prabhakar
- Karanjit Singh
- glob
- sqlite3
- pandas
- matplotlib
- numpy
- Generate the flat out files from TPC-E benchmark available at https://www.tpc.org/tpce/
- after generating the flat out make a directory named 'raw-data' in root folder
- execute
python3 src/extract.py
this will load the csv files into raw-data directory
- enter the SQLite3 shell type
sqlite3
in terminal - change the mode to csv type in the sqlite3 shell type
.mode=csv
- create tables
.read ./scripts/1_create_table.sql
- load data to the tables
.read ./scripts/load_data.sql
- create indexes
.read ./scripts/4_create_index.sql
- create foreign key indexes
.read ./scripts/4_create_fk_index.sql
- use
python3 ./src/transactions/trade_<transaction_name>_<transaction_type>.py
- Transactions diretory has four transactions trade_order,trade_update,trade_lookup and trade_status.
- Transactions directory also has inmemroy version of 4 transactions mentioned , these transactions creates database inmemory and executes frames in a single process.
- note logging should be turned on using
PRAGMA journal_mode=WAL;
- All the data(operations performed) will be stored .json files
- visualize the performance of different database configurations
python3 ./src/plot.py