Arrays, Records and Files

One dimensional arrays

An array is a numbered row of boxes that all hold the same type of data. In the exam you declare one with a fixed size, and the first box is number 0:

array scores[5]
scores[0] = 12
scores[1] = 18
scores[4] = 7

Python has no fixed-size array. A list does the same job and can also grow, so when you write an exam answer in Python, behave like an array: make the list the right size at the start, and do not use append or insert.

# array scores[5]
scores = [0] * 5
scores[0] = 12
scores[1] = 18
scores[4] = 7
print(scores)             # [12, 18, 0, 0, 7]
print(len(scores))        # 5

The positions of a five box array are 0, 1, 2, 3 and 4. Asking for a position that does not exist is an error:

# raises: IndexError
scores = [12, 18, 0, 0, 7]
print(scores[5])

Going through an array

Use the index when the position matters, and the value when it does not.

scores = [12, 18, 0, 0, 7]

for i in range(len(scores)):
    print("Position", i, "holds", scores[i])

for score in scores:
    print(score)

The array algorithms the exam keeps asking for

Every one of these is a loop with something carried from one pass to the next. Learn the shape of each.

marks = [56, 71, 38, 90, 64]

total = 0
for mark in marks:
    total = total + mark
average = total / len(marks)
print(total, average)                       # 319 63.8

highest = marks[0]
position = 0
for i in range(1, len(marks)):
    if marks[i] > highest:
        highest = marks[i]
        position = i
print(highest, "at position", position)     # 90 at position 3

passes = 0
for mark in marks:
    if mark >= 50:
        passes = passes + 1
print(passes, "passed")                     # 4 passed

Notice that highest starts as the first item, not as 0. Starting at 0 fails if every value is negative, and starting at the first item is always right.

Linear search

Check every item in turn. Return its position when found, and only say "not found" after the whole array has been checked.

def find(names, target):
    for i in range(len(names)):
        if names[i] == target:
            return i
    return -1

names = ["Asha", "Ben", "Chloe", "Dev"]
print(find(names, "Chloe"))     # 2
print(find(names, "Zoe"))       # -1

Parallel arrays

Two arrays can describe the same set of things, if the box at one position in one array belongs with the box at the same position in the other.

names = ["Asha", "Ben", "Chloe", "Dev"]
marks = [56, 71, 38, 90]

best = 0
for i in range(1, len(marks)):
    if marks[i] > marks[best]:
        best = i
print(names[best], "scored", marks[best])     # Dev scored 90

They only stay in step if every change is made to both. Sort one array on its own and the names no longer match the marks. That fragility is the reason records exist.

Two dimensional arrays

A two dimensional array is a table: rows and columns. OCR writes the row first and the column second, grid[1,2]. Python uses a list of lists, so the same box is grid[1][2].

array grid[3,4]
grid[1,2] = 9

# array grid[3,4]: 3 rows, 4 columns
grid = [[0] * 4 for row in range(3)]
grid[1][2] = 9
print(grid)               # [[0, 0, 0, 0], [0, 0, 9, 0], [0, 0, 0, 0]]

wrong = [[0] * 4] * 3
wrong[1][2] = 9
print(wrong)              # [[0, 0, 9, 0], [0, 0, 9, 0], [0, 0, 9, 0]]

Going through a table

An outer loop for the rows, and an inner loop for the columns. The inner loop runs all the way round for every single pass of the outer one.

marks = [
    [56, 61, 70],
    [72, 68, 75],
    [38, 45, 40],
    [90, 85, 94],
]

for row in range(len(marks)):
    total = 0
    for col in range(len(marks[row])):
        total = total + marks[row][col]
    print("Student", row, "average", round(total / 3, 1))

The total is set to 0 inside the outer loop, so each student starts afresh. Move that line above the outer loop and every student's total includes the one before.

To go down the columns instead, swap the loops, and the outer loop is now the column:

marks = [
    [56, 61, 70],
    [72, 68, 75],
    [38, 45, 40],
    [90, 85, 94],
]

for col in range(3):
    best = 0
    for row in range(4):
        if marks[row][col] > best:
            best = marks[row][col]
    print("Best in test", col, "is", best)

A board game

Noughts and crosses is a 3 by 3 array of characters. Checking a row is one comparison:

board = [["X", "X", "X"],
         ["O", "X", "O"],
         [" ", "O", " "]]

for row in range(3):
    if board[row][0] == board[row][1] == board[row][2] == "X":
        print("X wins on row", row)

Records

An array holds many items of one type. A record holds the facts about one thing, and those facts can be of different types. Each fact is a field.

record Student
    name : string
    age : integer
    passed : Boolean
endrecord

Python has no record type at this level. A dictionary does the job, with the field names as keys.

student = {"name": "Asha", "age": 15, "passed": True}
print(student["name"], student["age"])      # Asha 15
student["age"] = 16
print(student)             # {'name': 'Asha', 'age': 16, 'passed': True}

An array of records

A list of dictionaries is a table where each row is one record. This is the structure behind nearly every real program that stores people, products or scores.

students = [
    {"name": "Asha", "age": 15, "mark": 56},
    {"name": "Ben", "age": 16, "mark": 71},
    {"name": "Chloe", "age": 15, "mark": 38},
]

for s in students:
    print(s["name"], "is", s["age"], "and scored", s["mark"])

print(students[1]["name"])          # Ben, the second record

for s in students:
    if s["mark"] < 50:
        print(s["name"], "needs a resit")      # Chloe needs a resit

A class gives a record named, checked fields. It is more than the GCSE needs, but it is how larger programs do it:

from dataclasses import dataclass

@dataclass
class Student:
    name: str
    age: int
    passed: bool

asha = Student("Asha", 15, True)
print(asha.name, asha.age)      # Asha 15
print(asha)                     # Student(name='Asha', age=15, passed=True)

Why records beat parallel arrays

Text files

Variables vanish when a program ends. A file keeps data after it does. The exam uses text files: lines of characters, read and written one at a time.

writeLine adds the end of line for you. Python's write does not, so you add \n yourself.

Open modes

file = open("scores.txt", "w")
file.write("Asha,56\n")
file.write("Ben,71\n")
file.write("Chloe,38\n")
file.close()

file = open("scores.txt", "r")
for line in file:
    print(line.strip())
file.close()

Each line read from a file still has its \n on the end. strip() takes it off. Forget it and every printed line is followed by a blank one, and a number read this way carries a newline into every comparison.

Reading records from a file

Each line is one record and the commas separate the fields. split(",") cuts a line into a list, and everything read from a file is text, so a number must be cast.

file = open("scores.txt", "w")
file.write("Asha,56\nBen,71\nChloe,38\n")
file.close()

students = []
file = open("scores.txt")
for line in file:
    name, mark = line.strip().split(",")
    students.append({"name": name, "mark": int(mark)})
file.close()

total = 0
for s in students:
    total = total + s["mark"]
print(students[0])                    # {'name': 'Asha', 'mark': 56}
print("Average", total / len(students))      # Average 55.0

Closing for you

The with statement closes the file when its block ends, even if something goes wrong inside it, so you cannot forget.

with open("log.txt", "w") as file:
    file.write("first\n")

with open("log.txt", "a") as file:
    file.write("second\n")

with open("log.txt") as file:
    first = file.readline()
    rest = file.readlines()
print(repr(first), rest)           # 'first\n' ['second\n']

A file that is not there

try:
    file = open("missing.txt")
except FileNotFoundError:
    print("That file does not exist yet")

Key terms you must be able to define

Array

A fixed size, numbered collection of items that are all the same data type, indexed from 0.

Index

The number that identifies one item in an array. The first is 0.

Two dimensional array

An array of arrays, organised as rows and columns, accessed with two indexes.

Record

A collection of related fields, which may be different data types, describing one thing.

Field

One named item of data inside a record, such as name or age.

Parallel arrays

Two or more arrays where the items at the same index belong together.

Linear search

Checking each item in turn from the start until the one wanted is found or the end is reached.

Text file

A file of characters organised in lines, read and written by a program one line at a time.

Append mode

Opening a file so that new data is added to the end and existing data is kept.

Flashcards

Say the definition out loud before you flip the card. Retrieving it from memory is what builds recall, reading it again is not.

Reach the top grade on this topic

Marks on data structure questions are awarded for the shape of the loop and for the details that beginners drop.

• Start a maximum or minimum at the first item of the array, never at zero
• Reset a running total inside the outer loop of a two dimensional array
• Put the not found result after the loop, never inside it
• Cast every value read from a file, and strip the newline from every line
• Close every file you open, or use with, and say which mode you want and why

Mistakes that cost marks here

marks = [[56, 61, 70], [72, 68, 75], [38, 45, 40], [90, 85, 94], [60, 60, 62]]   # given in the question

for student in range(5):
    total = 0
    for test in range(3):
        total = total + marks[student][test]
    print(total)

file = open("scores.txt", "w")        # set up the file the question describes
file.write("56\n71\n38\n90\n64\n")
file.close()

file = open("scores.txt", "r")
highest = int(file.readline())
for line in file:
    number = int(line)
    if number > highest:
        highest = number
file.close()
print(highest)