Matching question type

Database tables

quiz_match

This table is only used in the code for saving matching questions and can therefore be considered redundant.

id

int(10) unsigned NOT NULL auto_increment,

Primary key

question

int(10) unsigned NOT NULL default '0',

Foreign key to the id field in the quiz_questions table

subquestions

varchar(255) NOT NULL default ,

a comma separated list of ids of the question-answer pairs used in this question. These ids refer to the id field of the quiz_match_sub table.

shuffleanswers

tinyint(4) NOT NULL default '1',

a flag that determines whether the answers should be shuffled, provided the quiz settings allows this.

quiz_match_sub

The quiz_match_sub extends the quiz_questions table. It stores the pairs of questions and answers (as strings) that need to be matched for a correct solution.

id

int(10) unsigned NOT NULL auto_increment,

code

int(10) unsigned NOT NULL default '0',

This field is randomly set to a unique integer. This code will be used to identify the response. The reason we do not simply use the id is that that would make it possible for students to look at the page source and match the answers and questions with the same id.

question

int(10) unsigned NOT NULL default '0',

Foreign key to the id field in the quiz_questions table

questiontext

text NOT NULL,

answertext

varchar(255) NOT NULL default ,

Response storage

The matching questiontype needs to store information about the order of its subquestions and the selected responses. Because each subquestion is identified by one record in the quiz_match_sub table, the ids of this table could be used to identify both the question and the response part. This is indeed how it was done until Moodle 1.5. This however had the drawback that students who looked at the page source could figure out which answer belonged to which question. Since Moodle 1.6 the response part is encoded with respect to the new 'code' field in the quiz_match_sub table. The entry in the 'code' field is randomly created with rand() when a subquestion is created.

For each subquestion a '-'-separated pair is created, giving first the id of the match_sub record that provided the subquestion and then the code of the match_sub record that provided the selected response. If no response is selected, the second entry in the pair is set to '0'. These pairs are then put into a comma separated list to allow each subquestion to store a response.

Question options object

The $questions->options object has the properties

subquestions

an array of all the records from the quiz_match_sub table for this question, i.e., all the question-answer pairs, indexed by the ids.

shuffleanswers

a flag that indicates whether the answers should be shuffled, provided this is enabled at the quiz level.

State options object

The $state->options object has a single property 'subquestions' which holds an array of objects representing the question-answer pairs, indexed by the ids. Each entry from the table is an object with the following properties:

id

question

questiontext

answertext

options

The options property in turn is an object with a single property answers which is an array of objects with the following properties:

id

the id of this subquestion

answer

the answertext for this subquestion

fraction

this is always set to 1 and has the effect that all subquestions carry the same weight.

The reason this is done in such a complicated manner is that it makes it easier to reuse the code for the matching question type also for the random short-answer questiontype.