spanoselias / TestingSQLCompliance

A framework for testing SQL compliance of current DBMSs

Implementation of a complete framework for evaluating the SQL-Compliance of current DBMSs

This project aims to build a complete framework for testing the SQL-Compliance of current DBMSs ( PostgreSQL, MySQL, IBM DB2, MS Server sql, Oracle Database). The core component of the framework is consisted by the random SQL generator tool, the comparison tool and the datafiller for generate random realistic data. The project is implemented in Java Programming Language. Moreover, this framework is implemented in such a way that a new DBMS can be added efficiently without the need to change the whole code.

Compilation & Run instructions using PostgreSQL

Linux

 Compile: javac -cp '.:postgresql-42.1.1.jar' SQLEngine.java
 Run: java -cp '.:postgresql-42.1.1.jar' SQLEngine

Windows

 Compile: javac -cp '.;postgresql-42.1.1.jar' SQLEngine.java
 Run: java -cp '.;postgresql-42.1.1.jar' SQLEngine

Note: The following parameters can be passed as arguments to the random query tool:

-maxTablesFrom, -maxAttrSel, -maxCondWhere, -maxAttrGrpBy, -probWhrConst,
-repAlias, -nestLevel, -arithCompSel, -distinct, -stringInSel,-stringInWhere, -rowcompar,
-DBMS, -user, -pass, -dbName, -isNULL,-isSelectAll

Example 1 of passing some parameters as arguments

 Compile: javac -cp '.:postgresql-42.1.1.jar' SQLEngine.java
 Run: java -cp '.:postgresql-42.1.1.jar' SQLEngine -pass testing1  -DBMS mysql -dbname  teststr -user root

Example 2

 javac -cp '.;postgresql-42.1.1.jar' SQLEngine.java -maxTablesFrom  5 -maxCondWhere 3 -maxAttrSel 4

Currently supported DBMSs:

 > MySql
 > Oracle
 > PostgreSQL
 > IBM DB2
 > MSSQL Server

Dependencies

> Linux or Windows
> Java 8 Update 92 or higher 
> Datafiller 
  

How to generate realistic data using datafiller

python datafiller.py --drop -n 0.33 schema.sql  > data.sql

Goals

The goals of this project are summarized as follow:

• Random Query Generator Tool

• Comparison Tool

• Datafiller Tool

• Highlight main differences among current DBMSs