Why database software is used to hold and manipulate data

(it applies to every section of Topics 17, 18 and 19)

• A 3-hour practical, 80 marks, one business, five questions in a fixed order. Question 1: fix the given table, build a data entry form, turn a flat extract into three related tables and import it. Question 2: a select query, an aggregate query and a grouped report. Question 3: a query with a calculated or counted field, then a chart. Question 4: a dashboard. Question 5: a 9-mark essay on changing the structure. A second 9-mark essay sits in Question 1 or Question 2.

• Six commands, each at a fixed tariff: Create (3 to 11 marks), Develop (11), Import (3), Set up (2), Identify and correct (1) and Analyse and recommend (9). There are no recall questions, no definitions and no multiple choice.

• Evidence: every task ends by screen printing a named view (Design, Form, Datasheet or Print Preview), pasting it into a word-processed candidate evidence template and re-saving the template. The two essays are typed into the same template. Only the template is handed in.

• Where Topic 17 earns marks: in the June papers of 2023, 2024 and 2025 it led one 9-mark essay (what information a user's task needs, what the database already holds and what is missing). Its ideas also carry roughly 14 marks a paper inside other tasks: clear labels and titles, helpful messages, sensible button captions, readable reports and charts.

• Timing: about 2.25 minutes a mark.

Every benefit above depends on the relational structure you met in Topic 5 (keys and relationships) and build in Topic 18 (data types, keys, referential integrity, validation).

A database that is one big table, like Greenlight's old sheet, loses most of them.

That is why "develop a relational structure" is worth 11 marks every year.

Two faults in Greenlight's May 2026 sheet:

• Learner Amira Khalil changed her mobile number. Her name and number were on 9 lesson rows. Owen changed 7 of them, so 2 later lessons still carried the old number and the reminder texts for those 2 lessons went to a phone she no longer used.

• Instructor Dev Sharma taught 46 one-hour lessons in May. The instructor column read "Dev Sharma" on 39 rows, "Dev" on 5 and "D Sharma" on 2. Farida filtered on "Dev Sharma" to work out his pay at £22 per lesson hour.

1. The first fault is redundancy: the number is held 9 times, so it must be changed 9 times. In a database it is held once, in Amira's tblLearner record, and each lesson finds it through LearnerID.

2. The second fault is inconsistent free text. The filter found 39 lessons, so Dev's pay was worked out as 39 × £22 = £858 instead of 46 × £22 = £1,012, which is £154 short (7 × £22).

3. In a database, the lesson stores InstructorID, chosen from a drop-down list of instructors, so the same instructor cannot be typed three ways.

Answer: both faults come from holding the data in one unstructured sheet.

A database holds each learner and instructor once and links lessons to them by key, so one change updates every lesson and every count is complete.

What it means: neither fault was anyone being careless in a way software cannot help.

The structure made the mistakes easy; a better structure makes them impossible.

Farida also uses a spreadsheet for her accounts.

Which tool suits each job?

Answer: a database is chosen when data about several related things must be kept, shared and kept accurate over time; a spreadsheet when the job is modelling numbers.

What it means: "a database can do calculations" is not a reason, because a spreadsheet can too.

The reasons are structure, integrity, sharing and querying across related tables.

Gloria Boateng, an instructor, wants a printed list of her lessons on 15 May 2026, earliest first, with each learner's name and mobile number so she can call if she is delayed.

1. The data sits in three tables: tblLesson (date, time, keys), tblLearner (names, mobile) and tblInstructor (Gloria's ID).

2. One select query joins them through their relationships and filters on two criteria:

3. The last two columns are criteria only, so Show is unticked.

Answer: the query returns Gloria's lessons for that day with names and numbers, in time order, and gives the right answer again tomorrow if the date is changed.

In the spreadsheet the same list meant filtering, copying and checking each number by hand.

What it means: this is what "manipulate" means here.

The full method for building queries, criteria and sorts is in Topic 19, "Retrieving data for specific purposes: queries"; here the point is why the database can do it: the data is held once, in related tables.

Task: Farida's cousin, who runs a small tutoring business, keeps all bookings in one spreadsheet and says a database "would just be a spreadsheet with extra steps".

Analyse the problems her single spreadsheet is likely to cause, and recommend how a database should be set up to avoid them.

(9)

Weaker answer (Level 1, about 2 marks): "Databases are better than spreadsheets because they are more secure and faster and can hold more data.

She should use Access because it is better for bookings."

Every claim is generic, none is tied to bookings data, and "use Access" names a product, not a solution.

Top answer (Level 3, about 8 marks):

• Analysis: each booking row repeats the student's name and contact details, so a changed phone number must be found and changed on every row, and missed rows leave old numbers in use. Tutor and subject names typed freely appear in several spellings, so filters and counts miss bookings. Nothing stops a booking for a tutor at a time they already teach, or a 300-minute session. Anyone with the file can overwrite or delete history, and two people editing copies of it end up with two versions.

• Recommendation: three tables, tblStudent (StudentID primary key, names, phone, email), tblTutor (TutorID primary key, names, subjects) and tblSession (TutorID and SessionDateTime as a composite primary key, StudentID as a foreign key, Minutes), linked one-to-many with referential integrity enforced. A range rule on Minutes with a helpful message; the composite key stops a tutor being booked twice at one time. Staff enter bookings on a form with drop-down lists of students and tutors, so names are chosen, not typed, and reach the tables only through forms on a dashboard.

What lifts it: every problem is tied to the kind of data held, and every recommendation is a structure that would work.

• Essays are typed, never screen printed. Name the actual tables, fields and keys.

• Practical tasks in this section (the query in Worked example 3, the validation rule in the questions) follow the evidence rules of Topics 18 and 19: Design view with every table, field and criterion readable; Datasheet view with columns wide enough to read.