Lesson – 5

File Handling – Binary Files

Most of the files that we see in our computer system are called binary files.
Examples
 Document files - .pdf, .doc, .xls etc.
 Image files - .png, .jpg, .gif, .bmp etc.
 Video files - .mp4, .3gp, .mkv, .avi etc.
 Audio files: .mp3, .wav, .mka, .aac etc.
 Database files: .mdb, .accde, .frm, .sqlite etc.
 Archive files: .zip, .rar, .iso, .7z etc.
 Executable files: .exe, .dll, .class etc.
All binary files follow a specific format. We can open some binary files in the
normal text editor but we can’t read the content present inside the file. That’s
because all the binary files will be encoded in the binary format, which can be
understood only by a computer or a machine.
In binary files, there is no delimiter to end a line. Since they are directly in the
form of binary, hence there is no need to translate them. That’s why these files
are easy and fast in working.
Binary file operations

The four major operations performed using a binary files are :

1. Inserting/Appending record in a binary file.
2. Read records from a binary file.
3. Search a record in a binary file.
4. Update a record in a binary file
Also, sometimes it is necessary to write and read an object like a list, dictionary
or instances of a class on to a file, then binary file is needed. Files that store
objects as some byte stream are called binary files.

Inserting or adding (appending) a record into a binary file requires importing

pickle module to write or read data form a binary file. To open files in binary
mode, when specifying a mode, add 'b' to it.

1
Python provides a special module called pickle module. Python pickle
module is used for serializing and de-serializing a Python object structure.
Pickling is a way to convert a python object (list, dict, etc.) into a character
stream.

Unpickling is the inverse operation, whereby a byte stream (from a binary file
or bytes-like object) is converted back into an object hierarchy(back to the
original structure).

First we need to import the module. It provides two main methods for the
purpose – dump() and load()

For creation of a binary file pickle.dump() is used. This will write objects to the
file which is opened in binary access mode.

Syntax of dump() method

dump(object,fileobject)

Examples

1. Write a program to write a list to a binary file.

def fileoperation():
import pickle
lst=[10,20,30,40]
f=open("list1.dat",'wb')
pickle.dump(lst,f)
print("List added to binary file")
f.close()

fileoperation()

Output

List added to binary file

2. Write a program to write a dictionary to a binary file.

#program to write a dictionary

2
 import pickle
d1={'Python':90,'Java':95,'C++':85}
f=open('d.dat','wb')
pickle.dump(d1,f)
f.close()

Once data is stored using dump(), it can be used for reading. For reading
data from the binary file, we have to use pickle.load() to read the object
from the pickle file.

Syntax of load()

object=load(fileobject)

Examples

1. Write a program to display the contents of the binary file list1.dat and
display the contents on the screen.

import pickle
f=open("list1.dat","rb")
l=pickle.load(f)
print(l)

Output

[10, 20, 30, 40]

2. Write a program to display the contents of the binary file d.dat and
display the contents on the screen.

import pickle
f=open('d.dat','rb')
d=pickle.load(f)
print(d)

Output

{'Python': 90, 'Java': 95, 'C++': 85}

3
3. Write a program to insert/append a record in the binary file "student.dat".

import pickle
record=[]
while True:
rollno=int(input("Enter roll number:"))
name=input("Enter name:")
marks=int(input("Enter marks:"))
data=[rollno,name,marks]
record.append(data)
ch=input("Want to enter more records?")
if ch.upper()=='N':
break

f=open("Student.dat","wb")
pickle.dump(record,f)
print("record added")
f.close()

Enter roll number:1

Enter name:Akash
Enter marks:45
Want to enter more records?y
Enter roll number:2
Enter name:Giny
Enter marks:67
Want to enter more records?y
Enter roll number:3
Enter name:john
Enter marks:78
Want to enter more records?N
record added

As shown in the above program, inserting and adding record in student file
begins with importing pickle module. Then, iterating for the number of records
to be added is done using while loop. Input is accepted from the user for roll
number, name and marks. These fetched values are saved as a list to be
appended. Once the record is created, it is appended to the binary file “student”
using the dump() method, which writes the object into the opened file. Finally

4
the file is closed explicitly and the record is added into the student.dat file as
shown in the output. Since binary file contains unreadable characters, it
becomes difficult to read and understand, but directly used and understood by
the computer.

4. Write a program to read a record from the binary file "Student.dat".

import pickle
f=open("Student.dat","rb")
stud_rec=pickle.load(f) # to read the object from the binary file
print("Contents of the file are:")
#reading the fields
for i in stud_rec:
rollno=i[0]
name=i[1]
marks=i[2]
print(rollno,name,marks)
f.close()

Output

Contents of the file are:

1 Akash 45
2 Giny 67
3 john 78

Searching a record in a binary file

Searching the binary file "student" is carried out on the basis of the roll number
entered by the user. The file is opened in the read-binary mode and gets stored
in the file object, f. load() method is used to read the object from the opened
file. A variable ‘found’ is used which will tell the status of the search operation
being successful or unsuccessful. Each record from the file is read and the
contents of the field, roll no, is compared with the roll number to be searched.
Upon the search being successful, appropriate message is printed.

5. Write a program to search for a record in the binary file “Student.dat”

import pickle
f=open("Student.dat","rb")

5
 stud_rec=pickle.load(f) # to read the object from the binary file
found=0
rno=int(input("Enter the roll number to search for:"))

for i in stud_rec:
if i[0]==rno:
print("Search Successful",i[1],"found")
found=1
break
if found==0:
print("Record not found")
f.close()

Output

Enter the roll number to search for:3

Search Successful john found

Note

 i[0] stands for roll number of the first student

 i[1] stands for name of the first student
 i[2] stands for the marks of the first student

Updating a record in a binary file

Updating a record in the file requires roll number to be fetched from the user
whose name is to be updated.

6. Write a program to update a record in the file. Enter the roll number of
the student whose name has to be updated.

import pickle
f=open("Student.dat","rb+")
stud_rec=pickle.load(f) # to read the object from the binary file
found=0
rno=int(input("Enter the roll number to search for:"))

for i in stud_rec:
if i[0]==rno:

6
 print("Current name is:",i[1])
i[1]=input("Enter the new name:")
found=1
break
if found==0:
print("Record not found")
else:
f.seek(0) # taking the file pointer to the beginning of the file
# program will not work without this statement
pickle.dump(stud_rec,f)
print("Name updated!")

f.close()

Output

Enter the roll number to search for:1

Current name is: Akash
Enter the new name:Don
Name updated!

Now if the file contents are displayed the output will be

Contents of the file are:

1 Don 45
2 Giny 67
3 john 78

Once the record is found, the file pointer is moved to the beginning of the file
using seek(0) statement, and then the changed name is written to the file and the
record is updated. seek() method is used for random access to the file.

7. Write a program to delete a record from the file “Student.dat”.

8. Write a menu-driven program to perform all the basic operations using

dictionary on student binary file such as inserting, reading, updating,
searching and deleting a record.

#p6.py
#Accepting data for a dictionary

7
def insertrec():
rollno=int(input("Enter the roll number:"))
name=input("Enter the name:")
marks=int(input("Enter marks:"))
#Creating the dictionary
rec={"Rollno":rollno,"Name":name,"Marks":marks}
#Writing to the dictionary
f=open("Stud.dat","ab")
pickle.dump(rec,f)
f.close()

#Reading the records

def readrec():
f=open("stud.dat","rb")
while True:
try:
rec=pickle.load(f)
print("Roll no:",rec['Rollno'])
print("Name:",rec['Name'])
print("Marks:",rec['Marks'])
except EOFError:
break
f.close()

#Searching for a record based on roll number

def searchrollno(r):
f=open("stud.dat","rb")
flag=False
while True:
try:
rec=pickle.load(f)
if rec['Rollno']==r:
print("Roll no:",rec['Rollno'])
print("Name:",rec['Name'])
print("Marks:",rec['Marks'])
flag=True
except EOFError:
break
if flag==False:
print("No record found")

8
 f.close()
#Mark modification for a roll number
def updatemarks(r,m):
f=open("stud.dat","rb")
reclist=[]
while True:
try:
rec=pickle.load(f)
reclist.append(rec)
except EOFError:
break
f.close()
for i in range(len(reclist)):
if reclist[i]['Rollno']==r:
reclist[i]['Marks']=m
f=open("stud.dat",'wb')
for x in reclist:
pickle.dump(x,f)
f.close()
def deleterec():
f=open("stud.dat","rb")
reclist=[]
while True:
try:
rec=pickle.load(f)
reclist.append(rec)
except EOFError:
break
f.close()
f=open("stud.dat","wb")
for x in reclist:
if x['Rollno']==r:
continue
pickle.dump(x,f)
f.close()
#__main__
import pickle
while True:
print("1.Insert Record")
print("2.Display Record")

9
 print("3.Search Record")
print("4.Update Record")
print("5.Delete Record")
print("0.Exit")
print("Enter your choice:")
ch=int(input("Enter choice:"))
if ch==0:
break
elif ch==1:
insertrec()
elif ch==2:
readrec()
elif ch==3:
r=int(input("Enter a roll number to search:"))
searchrollno(r)
elif ch==4:
r=int(input("Enter a roll number to update:"))
m=int(input("Enter the new marks:"))
updatemarks(r,m)
elif ch==5:
r=int(input("Enter a roll number to delete:"))
deleterec()

Random access in files using tell() and seek()

Till now in all the programs, sequential processing of data in a text and binary
file was emphasized. But files in Python allow random access of the data as
well using built-in methods seek() and tell().

seek() - seek() function is used to change the position of the file handle (file
pointer) to a given specific position. File pointer is like a cursor, which defines
from where the data has to be read or written in the file. Python file method
seek() sets the file’s current position at the offset.

This argument is optional and defaults to 0, which means absolute file

positioning. Other values are: 1, which signifies seek is relative (may change) to
the current position and 2 means seek is relative to the end of file. There is no
return value. The reference point is defined by the "from_what" argument.

It can have any of the three values:

10
0: sets the reference point at the beginning of the file, which is by default.
1: sets the reference point at the current file position.
2: sets the reference point at the end of the file.

seek() can be done in two ways:

• Absolute Positioning
• Relative Positioning

Absolute referencing using seek() gives the file number on which the file
pointer has to position itself.

Syntax for seek()    

f.seek(file_location)  #where f is the file pointer

For example, f.seek(20) will give the position or file number where the
file pointer has been placed. This statement shall move the file pointer to
20th byte in the file no matter where the pointer is.

Relative referencing/positioning has two arguments, offset and the

position from which it has to traverse.

Syntax for relative referencing is:

f.seek(offset,from_what)  #where f is file pointer

Example

f.seek(-10,1) from current position, move 10 bytes backward

f.seek(10,1) from current position, move 10 bytes forward
f.seek(-20,1) from current position, move 20 bytes backward
f.seek(10,0) from beginning of file, move 10 bytes forward

Note

 Python 3.x only supports text file seeks from the beginning of the file.
 seek() with negative offset only works when the file is opened in binary
mode.

11
tell() - tell() returns the current position of the file read/write pointer within the
file.

Syntax

f.tell()      #where f is file pointer

 When you open a file in reading/writing mode, the file pointer rests at 0th
byte.
 When you open a file in append mode, the file pointer rests at last byte.

Example-1

The contents of test.txt are:

f=open("test.txt")
print("Before reading",f.tell())
a=f.read()
print("After reading",f.tell())
f.seek(0)
#brings the file to the 0th byte of the file
print("From beginning again",f.tell())
s=f.read(4)
print("First 4 bytes are:",s)
print(f.tell())
s=f.read(3)
print("Next 3 bytes are:",s)
print(f.tell())
f.close()

Output

12
 Before reading 0
After reading 35
From beginning again 0
First 4 bytes are: This
4
Next 3 bytes are: is
7

Example-2

f=open("File.txt",'w')
f.write("This is a test\n")
f.write("program in\n")
f.write("Python")
f.close()

f=open("File.txt",'r')
f.seek(10) # by default 0
print(f.tell())
print(f.readline())
f.close()

OR

f=open("File.txt",'r')
f.seek(10,0)
print(f.tell())
print(f.readline())
f.close()

Output

10
test

Note

All the remaining characters from the 10th byte will be displayed until end
of line character.

13
Example-3
f=open("File.txt",'w')
f.write("This is a test") # ‘\n’ removed
f.write("program in\n")
f.write("Python")
f.close()

f=open("File.txt",'r')
f.seek(10)
print(f.tell())
print(f.readline())
f.close()

Output

10
testprogram in

Note

If there is no ‘\n’ in the first line then all the remaining characters from
the 10th byte will be displayed until ‘\n’ will be displayed.

Example-4

f=open("File.txt",'w+')
f.write("This is a test \n")
f.write("program in\n")
f.write("Python")

# Second parameter is by default 0

# sets Reference point to twentieth
# index position from the beginning
f.seek(20)

# prints current position

print(f.tell())

print(f.readline())
f.close()

14
Output

20
gram in

Example-5

f=open("File.txt",'w')
f.write("This is a test\n") # ’\n’ included
f.write("program in\n")
f.write("Python")
f.close()

f=open("File.txt",'r')
f.seek(10)
print(f.tell())
f.seek(5,0)
print(f.tell())
print(f.readline())
f.close()

Output

10
5
is a test

Note

Arguments 1 and 2 are not supported by text files.

Example-6

f=open("File.txt",'r')
f.seek(0,1)
print(f.tell())
print(f.readline())
f.close()

15
Output
0
This is a test

Note

Non-zero seek() which works with text files f.seek(0,1)

Non-zero seek() which works with text files f.seek(0,2)

Example-7

f=open("File.txt",'rb')
f.seek(0,1)
print(f.tell())
print(f.readline())
f.close()

Output

b'This is a test\r\n'

Example-8

f=open("File.txt",'r')
f.seek(0,2)
print(f.tell())
print(f.readline())
f.close()

Output

34

Example-9

f=open("File.txt",'r')
f.seek(20)
f.seek(10)
print(f.tell())
f.seek(0,1)

16
 print(f.tell())
f.seek(5)
print(f.tell())
print(f.readline())
f.close()

Output

10
10
5
is a test

Example-10

f=open("File.txt",'rb')
f.seek(10)
print(f.tell())
print(f.readline())
f.seek(5,1)
print(f.tell())
print(f.readline())
f.seek(5,1)
print(f.tell())
print(f.readline())
f.close()

Output

10
b'test\r\n'
21
b'am in\r\n'
33
b'n'

Example-11

f=open("File.txt",'w')
f.write("This is a test\n")

17
 f.write("program in\n")
f.write("Python")
f.close()

f=open("File.txt",'rb')
# sets Reference point to 30th
# position to the left from end
f.seek(-30, 2)

# prints current position

print(f.tell())

# Converting binary to string and

# printing
print(f.readline())

f.close()

Output

4
b' is a test\r\n'

Example-12

f=open("File.txt",'w')
f.write("This is a test\n")
f.write("program in\n")
f.write("Python")
f.close()

f=open("File.txt",'rb')
# sets Reference point to tenth
# position to the left from end
f.seek(-10, 2)

# prints current position

print(f.tell())

# Converting binary to string and

# printing
18
 print(f.readline())

f.close()

Output
24
b'in\r\n'

Note: Reference point at current position / end of file cannot be set in text mode
except when offset is equal to 0.

19