Computer Science · Class 12 · CSV Files

CSV filespractice⏱️ 20 min read

Practice Programs

Eight programs, and no new syntax in any of them. Every one is csv.writer() or csv.reader() with something ordinary in the loop — a comparison, a counter, a list being built. Read them for the shape, not for the lines.

The lesson these programs practisereader() — the lesson most of these programs lean on →

They all work on one file

Program 1 creates students.csv and everything after it reads or changes that same file. The transcripts below were run in order, so the file each program sees is whatever the one before it left behind.

students.csv

Roll,Name,Marks
1,Ravi,78
2,Meera,91
3,Amit,65

11 · Create a CSV from what the user types

📋 The problem

Write a program that asks for a roll number, a name and marks, writes each record into students.csv, and keeps asking until the user says no.

create_csv.py

import csv

f = open('students.csv', 'w', newline='')
w = csv.writer(f)
w.writerow(['Roll', 'Name', 'Marks'])

more = 'y'
while more == 'y':
    roll = int(input('Roll number : '))
    name = input('Name        : ')
    marks = int(input('Marks       : '))
    w.writerow([roll, name, marks])
    more = input('One more (y/n)? ')

f.close()
print('students.csv is ready.')

Output

Roll number : 1
Name        : Ravi
Marks       : 78
One more (y/n)? y
Roll number : 2
Name        : Meera
Marks       : 91
One more (y/n)? y
Roll number : 3
Name        : Amit
Marks       : 65
One more (y/n)? n
students.csv is ready.

w.writerow(['Roll', 'Name', 'Marks'])

The header, above the loop, so it is written once.

while more == 'y':

A while loop, because nobody knows in advance how many records there will be. more starts as 'y' so the loop runs at least once.

int(input(...))

The int() is not needed for the file — writerow() would take the string quite happily. It is there because a roll number is a number, and the habit is worth keeping.

f.close()

Outside the loop. Close it once, when there is nothing more to add.

22 · Display every record, and count them

📋 The problem

Display the contents of students.csv as a neat table, and print how many records it holds.

display_csv.py

import csv

f = open('students.csv', 'r', newline='')
r = csv.reader(f)

header = next(r)
print(header[0], header[1], header[2], sep='\t')
print('-' * 24)

count = 0
for row in r:
    print(row[0], row[1], row[2], sep='\t')
    count = count + 1

f.close()

print('-' * 24)
print('Records:', count)

Output

Roll	Name	Marks
------------------------
1	Ravi	78
2	Meera	91
3	Amit	65
------------------------
Records: 3

Three records, four rows

count is 3 because next(r) took the header out before the loop started. Count without skipping it and you report 4 — the single most common mistake in this chapter's exam answers.

33 · Search for one record

📋 The problem

Ask for a roll number and display that student's name and marks. If there is no such student, say so.

search_csv.py

import csv

wanted = int(input('Roll number to find: '))

f = open('students.csv', 'r', newline='')
r = csv.reader(f)
next(r)

found = False
for row in r:
    if int(row[0]) == wanted:
        print('Name :', row[1])
        print('Marks:', row[2])
        found = True
        break

f.close()

if found == False:
    print('No student with roll number', wanted)

Output

Roll number to find: 2
Name : Meera
Marks: 91

the same program, asked for a roll number that is not there

Output

Roll number to find: 7
No student with roll number 7

int(row[0]) == wanted

row[0] is the string '2' and wanted is the number 2, and those are never equal. One side has to be converted — here the row is, because that is the value that came out of the file.

found = False

Set before the loop. Without it there is no way to tell 'nothing matched' from 'the loop has not run yet'.

break

Roll numbers are unique, so there is nothing to gain from reading the rest of the file.

if found == False:

Outside the loop, and after f.close(). Inside the loop it would print the message once for every row that did not match.

44 · Add a record to the end

📋 The problem

Ask for one new student's details and add them to students.csv without disturbing the records already there.

append_csv.py

import csv

roll = int(input('Roll number : '))
name = input('Name        : ')
marks = int(input('Marks       : '))

f = open('students.csv', 'a', newline='')
w = csv.writer(f)
w.writerow([roll, name, marks])
f.close()

print('Record added.')

Output

Roll number : 4
Name        : Sneha
Marks       : 88
Record added.

students.csv — afterwards

Roll,Name,Marks
1,Ravi,78
2,Meera,91
3,Amit,65
4,Sneha,88

Key Takeaway

'a', and no header. The file already has one, and writing another would put Roll,Name,Marks in the middle of the data as if it were a student. Change that one letter to 'w' and the other three records are gone before Sneha is written.

55 · Find the topper, and count who passed 75

📋 The problem

Print the name and marks of the student with the highest marks, and how many students scored more than 75.

topper_csv.py

import csv

f = open('students.csv', 'r', newline='')
r = csv.reader(f)
next(r)

best_name = ''
best_marks = 0
above = 0

for row in r:
    marks = int(row[2])
    if marks > best_marks:
        best_marks = marks
        best_name = row[1]
    if marks > 75:
        above = above + 1

f.close()

print('Topper       :', best_name, 'with', best_marks)
print('Above 75     :', above)

Output

Topper       : Meera with 91
Above 75     : 3

marks = int(row[2])

Converted once, at the top of the loop, and used twice below. Without the int() the comparison compares strings, and '9' > '75' comes out True because it is judged letter by letter.

best_marks = 0

Starting at 0 works because marks are never negative. Starting at the first record's marks is the safer habit for data that could be.

one loop, two questions

The file is read once and both answers come out of the same pass. Opening it twice would work and would be twice the work.

66 · Copy some records into a second file

📋 The problem

Copy every student who scored 80 or more into toppers.csv, keeping the header, and report how many were copied.

copy_csv.py

import csv

fin = open('students.csv', 'r', newline='')
r = csv.reader(fin)
header = next(r)

fout = open('toppers.csv', 'w', newline='')
w = csv.writer(fout)
w.writerow(header)

copied = 0
for row in r:
    if int(row[2]) >= 80:
        w.writerow(row)
        copied = copied + 1

fin.close()
fout.close()

print(copied, 'records copied into toppers.csv')
print(open('toppers.csv', newline='').read())

Output

2 records copied into toppers.csv
Roll,Name,Marks
2,Meera,91
4,Sneha,88

Two files open at once is normal

Two open() calls, two variables, two close() calls. Name them for their jobs — fin and fout, or f1 and f2 — because using f twice makes the first file impossible to close.

Note w.writerow(header): the header row was read out of the first file and written straight into the second, so the copy is a proper CSV rather than a heap of records.

77 · Change one student's marks

📋 The problem

Ask for a roll number and new marks, and update that record in students.csv.

There is no way to change one line of a file in place — the new record is rarely the same length as the old one, and everything after it would have to shift. So an update is always three steps, exactly as it was for text files and binary files:

read it all in

list(csv.reader(f)) — every row, in a list, in memory.

change the list

Ordinary Python. Find the row, assign the new value.

write it all back

Open in 'w' and writerows() the whole list.

update_csv.py

import csv

wanted = int(input('Roll number to update: '))
new_marks = int(input('New marks           : '))

# 1. read everything into memory
f = open('students.csv', 'r', newline='')
rows = list(csv.reader(f))
f.close()

# 2. change the one row
found = False
for row in rows:
    if row[0] != 'Roll' and int(row[0]) == wanted:
        row[2] = new_marks
        found = True

# 3. write them all back
if found == True:
    f = open('students.csv', 'w', newline='')
    w = csv.writer(f)
    w.writerows(rows)
    f.close()
    print('Updated.')
else:
    print('No such roll number.')

Output

Roll number to update: 3
New marks           : 72
Updated.

students.csv — afterwards

Roll,Name,Marks
1,Ravi,78
2,Meera,91
3,Amit,72
4,Sneha,88

rows = list(csv.reader(f))

The header is in this list too — it is row zero. That is why it is written back at the end without any special handling.

row[0] != 'Roll'

The guard that keeps int('Roll') from raising ValueError. Because the header is in the list, every loop over rows has to step around it.

row[2] = new_marks

row is a list, and lists are mutable — changing it here changes the copy inside rows. This is the whole reason step 1 read into a list rather than looping the file.

if found == True:

The file is only rewritten when something actually changed. Without this check, a wrong roll number rewrites the file for no reason — harmless here, but the habit matters.

Close the file before you reopen it

Step 1 closes the file before step 3 opens it again in 'w'. They are two different handles on the same file, and the reading one must be finished with before the writing one empties it.

88 · Delete a record

📋 The problem

Ask for a roll number and remove that student from students.csv.

The same three steps. The only difference is step 2: instead of changing a row, you build a new list of the rows you are keeping.

delete_csv.py

import csv

wanted = int(input('Roll number to delete: '))

f = open('students.csv', 'r', newline='')
rows = list(csv.reader(f))
f.close()

kept = []
for row in rows:
    if row[0] == 'Roll' or int(row[0]) != wanted:
        kept.append(row)

f = open('students.csv', 'w', newline='')
w = csv.writer(f)
w.writerows(kept)
f.close()

print(len(rows) - len(kept), 'record deleted')

Output

Roll number to delete: 1
1 record deleted

students.csv — afterwards

Roll,Name,Marks
2,Meera,91
3,Amit,72
4,Sneha,88

Key Takeaway

Keep, do not remove. Deleting items out of a list while you are looping over it makes the loop skip rows, and it is a bug that only shows up when two records in a row need deleting. Building a second list avoids the whole problem, and len(rows) - len(kept) counts what went.

The row[0] == 'Roll' half of the condition is what keeps the header. Written as or, it is checked first — so on the header row the int(row[0]) on the other side never runs, and never raises ValueError.

9Run one yourself

Everything on this page in a single program: build the file, read it back, search it, and change a record.

all_of_it.py

import csv

# create
students = [[1, 'Ravi', 78], [2, 'Meera', 91], [3, 'Amit', 65]]
f = open('students.csv', 'w', newline='')
w = csv.writer(f)
w.writerow(['Roll', 'Name', 'Marks'])
w.writerows(students)
f.close()

# display
f = open('students.csv', 'r', newline='')
r = csv.reader(f)
next(r)
for row in r:
    print(row[0], row[1], row[2], sep='\t')
f.close()

# update roll 3 to 72
f = open('students.csv', 'r', newline='')
rows = list(csv.reader(f))
f.close()

for row in rows:
    if row[0] != 'Roll' and int(row[0]) == 3:
        row[2] = 72

f = open('students.csv', 'w', newline='')
csv.writer(f).writerows(rows)
f.close()

print()
print(open('students.csv', newline='').read())

import csv

# create
students = [[1, 'Ravi', 78], [2, 'Meera', 91], [3, 'Amit', 65]]
f = open('students.csv', 'w', newline='')
w = csv.writer(f)
w.writerow(['Roll', 'Name', 'Marks'])
w.writerows(students)
f.close()

# display
f = open('students.csv', 'r', newline='')
r = csv.reader(f)
next(r)
for row in r:
    print(row[0], row[1], row[2], sep='\t')
f.close()

# update roll 3 to 72
f = open('students.csv', 'r', newline='')
rows = list(csv.reader(f))
f.close()

for row in rows:
    if row[0] != 'Roll' and int(row[0]) == 3:
        row[2] = 72

f = open('students.csv', 'w', newline='')
csv.writer(f).writerows(rows)
f.close()

print()
print(open('students.csv', newline='').read())

10The patterns behind all eight

Counting anything

count = 0 before the loop, count = count + 1 inside it under an if. Records, passes, failures — the same three lines.

Searching

found = False before the loop, set it True and break on a match, report the failure after the loop.

Best or worst

Hold the best so far in a variable, compare each row against it, replace when the row beats it.

Filtering into a new file

Two files open at once. Read a row, test it, and writerow() the ones that pass.

Changing or deleting

Read all → change the list → write all back. Never try to edit a line where it sits.

The header, every time

next(r) when reading, writerow() once when creating, nothing at all when appending — and a guard when looping over list(csv.reader(f)).

✍️ Now write these yourself

1
Count how many students scored below 40 and print their names.
Hint · Program 5's shape, with the comparison turned round.
2
Print the average marks of the whole class.
Hint · Total in one variable and a count in another, then divide. int(row[2]) both times.
3
Search by name instead of roll number, and handle two students with the same name.
Hint · No break, and count the matches instead of stopping at the first.
4
Write a program that adds 5 marks to every student, up to a maximum of 100.
Hint · Read all, loop the list changing row[2], write all back.
5
Merge two CSV files with the same columns into a third, with one header at the top.
Hint · next() on both readers; writerow() the header once.

Quick Check

Why does a search program need found = False before the loop?

Quick Check

Which mode adds a record to an existing CSV without losing the others?

Quick Check

How do you change one record in a CSV file?

Quick Check

When looping over rows = list(csv.reader(f)), why does the code check row[0] != 'Roll'?

← Previous

🔀 Choosing the Separator

What is a Data Structure? 🧱