- More documentation in the root directory with a file for each DBMS.
- Maven support (Based on https://github.com/andl/benchmarkSQL/blob/master/pom.xml)
- Code formatted, and imports organized.
- New package structure and directories.
- Correction of comments for javadoc.
- FIXMEs in the code to include any other RDBMS.
- Scripts for Windows.
- Directory per RDBMS.
- Shell scripts for each RDBMS.
- Updated to log4j 2.
- Fixed one more preparedStatement() leak. Hopefully with the help of Oracle's V$OPEN_CURSOR view we got them all now.
- Fixed a possible deadlock problem in the NEW_ORDER transaction. Multiple parallel transaction could attempt to lock the same STOCK rows in reverse order. Sorting the order lines by item ID avoids this problem.
- Changed the status line to update only once per second. The previous implementation was getting rather noisy at high throughput.
- Fixed two preparedStatement() leaks that could cause ORA-01000 errors on longer runs with high throughput.
- Fixed a problem in the calculation of sleep time between transactions when using limitTxnsPerMin that could cause the test to hang at the end.
- Added support for escaping ; as ; in SQL files to be able to load functions and execute anonymous PL blocks (needed for next item).
- Changed the definition of history.hist_id into a plain integer with no special functionality. Two new database vendor specific SQL scripts allow to enable the column after data load as an auto incrementing primary key. See HOW-TO-RUN.txt for details.
- Upgrade to using JDK 7
- Upgrade to PostgreSQL JDBC 4.1 version 1101 driver
- Stop claiming to support DB2 (only Postgres & Oracle are well tested)
- Incorporate new PostgreSQL JDBC 4 version 1100 driver
- Changed default user from postgres to benchmarksql
- Added id column as primary key to history table
- Renamed schema to benchmarksql
- Changed log4j format to be more readable
- Created the "benchmark" schema to contain all tables
- Incorporate new PostgreSQL JDBC4 version 1003 driver
- Transaction rate pacing mechanism
- Correct error with loading customer table from csv file
- Status line report dynamically shown on terminal
- Fix lookup by name in PaymentStatus and Delivery Transactions (in order to be more compatible with the TPC-C spec)
- Rationalized the variable naming in the input parameter files (now that the GUI is gone, variable names still make sense)
- Default log4j settings only writes to file (not terminal)
- Removed Swing & AWT GUI so that this program is runnable from the command line
- Remove log4j usage from runSQL & runLoader (only used now for the actual running of the Benchmark)
- Fix truncation problem with customer.csv file
- Comment out "BadCredit" business logic that was not working and throwing stack traces
- Fix log4j messages to always show the terminal name
- Remove bogus log4j messages
- Config log4j for rotating log files once per minute
- Default flat file location to '/tmp/csv/' in table copies script
- Drop incomplete & untested Windoze '.bat' scripts
- Standardize logging with log4j
- Improve Logging with meaningful DEBUG and INFO levels
- Simplify "build.xml" to eliminate nbproject dependency
- Defaults read in from propeerties
- Groudwork laid to eliminate the GUI
- Default GUI console to PostgreSQL and 10 Warehouses
- Default build is now with JDK 1.6 and JDBC 4 Postgres 9.2 driver
- Remove outdated JDBC 3 drivers (for JDK 1.5). You can run as before by a JDBC4 driver from any supported vendor.
- Remove ExecJDBC warning about trying to rollback when in autocommit mode
- Remove the extraneous COMMIT statements from the DDL scripts since ExecJDBC runs in autocommit mode
- Fix the version number displayed in the console
- Dare to Compare
- Forked from the jTPCC project
Version 1.2.0 is a major update of jTPCC with the following highlights:
- Compliance with TPC-C Specification 5.4
- New module for database population
- Support for Oracle, MySQL and PostgreSQL
- Graphics plotted in the interface
- Dump the database into a file after population (for fast restore)
Check the Change Log or get the full version (including all required libraries and also drivers for MySQL and PostgreSQL).
This release considers current v1.0.7 features stable. Text output was optimized and there is no longer one tab per terminal to improve usability.
A small number of terminals caused text output to become very heavy on the processor after displaying many text lines.
Just a few changes to the interface, nothing new.
Some improvements to the interface and a better limit on the memory used by the output areas.