A tool to help in assessing student-submitted SQL.
Go to file
Andrew Lalis 9d87837813 Updated readme. 2019-03-05 11:50:21 +01:00
out/artifacts/SQL_Assesser_jar Updated readme. 2019-03-05 11:50:21 +01:00
resources Added EVERYTHING 2019-02-21 19:43:02 +01:00
src Fixed bug where first row of query result is skipped, added jar file to git. 2019-03-03 14:12:30 +01:00
.gitignore Updated readme. 2019-03-05 11:50:21 +01:00
README.md Updated readme. 2019-03-05 11:50:21 +01:00

README.md

SQL-Assesser

A tool to help in assessing student-submitted SQL.

Program Screenshot

Instructions for Use

To use this tool to assess and compare SQL, please ensure the following prerequisites are met:

  • A PostgreSQL database server is currently running (which you have full CREATEDB permissions to).
  • You know the hostname, port, username, and password which are needed to connect to this database.

The aforementioned information can be entered into the relevant fields in the user interface as shown in the image. Once this is done, you may begin comparing SQL. To do this, do the following:

  1. Load any initialization statements into the box titled Initialization. Anything in this box will be executed on both databases to prepare them for the assessment.
  2. Load the template statements (the answers, or confirmed correct statements) into the box titled Template. These will be executed on only one of the databases.
  3. Load the testing statements (things you want to test for correctness) into the box titled Testing. These will be executed on the other database.
  4. Click Execute to begin the assessment process. It is as follows:
    1. Delete and create two new empty testing databases.
    2. Run the Initialization code on both databases.
    3. Run the Template code on one database.
    4. Run the Testing code on the other database.
    5. Output the result of each SQL statement, plus an evaluation of the result of Testing when compared with Template. This is true if the same schema update and data manipulation statements were used, as well as the same result sets from select queries.

Notes

There are cases where the assesser will return a false negative (i.e. false even though the SQL works) because of the improbability of validating the order and content of statements other than SELECT.

If you think you can figure out how to account for things such as out-of-order insertions, updates, and deletions, and drawing transaction boundaries around related sets of statements that have common pre- and post-conditions, feel free to create a pull-request with the changes.