3/13/2012

ESc101: Variable Types

The content of these slides are taken from the lecture
slides of Prof. Arnab Bhattacharya, Prof. R.K. Ghosh,
Instructor: Krithika Venkataramani Prof. Dheeraj Sanghi and Prof. Manindra Agrawal
Semester 2, 2011-2012

1 2
Krithika Venkataramani ([email protected]) Krithika Venkataramani ([email protected])

Variables Types of Variables

 Variables signify data that may be modified  Each variable ‘type’ represents the domain of values
 Name of a variable can contain letters, digits and underscore _  Integer: int or char
 Example: i, y2k, big_name, bigger_name_2  Character: char
 Boolean: int or char
 Case-sensitive: camel, CAMEL and CaMeL are different
 real: double or float
 Name cannot start with a digit
 However they can store only a subset of the domain
However,
 Example: 1d is not valid
 int can store numbers from -231 to 231 -1
 Name can start with an underscore, but do not do so
 Initial values of variables are specified as constants of the
 Example: avoid valid names such as _bad
same type
 Certain keywords are special
 int i = 0;
 They are reserved and cannot be used
 double d = 1.4;
 Example: main, if
 char c = 'A'

3 4
Krithika Venkataramani ([email protected]) Krithika Venkataramani ([email protected])

1
 3/13/2012

Example Program to Add two numbers Different Data Types

/* Program to add two numbers */  Different types are needed because one type is not suitable
#include <stdio.h> // Include headers for representing data of another type.
void main() // Main function  Mixing types may result in precision loss, overflow,
{ underflow
int a=2, b=3, c; // Declare variables  Application performance suffers while performing
scanf("%d",
f("%d" &&a);) // Read
R d 'a'
' ' ffrom kkeyboard
b d numerically intensive computation if inappropriate data types
are used.
scanf("%d", &b); // Read 'b' from keyboard  Exceptions must be handled explicitly or they lead to errors.
c = a + b;  Use of appropriate type is important both for efficiency and
correctness
printf("%d\n", c); // Write 'c' to screen
}

5 6
Krithika Venkataramani ([email protected]) Krithika Venkataramani ([email protected])

Integer Types Integer Representation

 Two different int types: signed and unsigned  We represent integers as binary numbers
 Maximum signed int in 16 bit: 011111111111111, i,e., 215 - 1  For example: 56 will be stored as
 00111000
 Maximum unsigned int in 16 bit: 111111111111111, i.e., 216 - 1
 Possible types to suit our needs are:  How do we store negative numbers?
 1st bit of memory is usually the sign bit
 short int, unsigned short int, unsigned int, long int, unsigned long int.
 In case of 8 bit memory
memory, only 7 bits are for magnitude
magnitude.
 Similarly 32-bit memory would have 31 bits for magnitude
 Hence the largest positive integer that can be stored in an integer
variable on our PCs is: 231 – 1.
 Smallest number:
 There are variations on storing magnitude
 Overflow: Trying to store numbers outside the range

7 8
Krithika Venkataramani ([email protected]) Krithika Venkataramani ([email protected])

2
 3/13/2012

Different Real Number Types Representation of Real Numbers

 Real numbers of arbitrary precision cannot be represented  Use the scientific notation: f * bk
 Different types: float, double, long double  With this notation, we need to store f and k.
 double is more accurate than float  We also need to decide the value of ‘b’.
 1/3 is printed as 0.33333334326.. as a float, but 0.33333333333.. as a  The most commonly used representation is:
double
 Use 1 bit for sign
 double
d bl iis usedd ffor precision
i i critical
iti l calculations
l l ti  Value of b is taken as 2
 By default floating point constants are stored as a double.  Use 8 bits to store k (called exponent)
 To force float constant should be suffixed with f, i.e., 7.5f or 7.5F.  Use 23 bits to store f (called mantissa), in normalized form with
integer part of the fraction to be exactly 1 (e.g. 1.0011)
 Format specifier "%lf", "%Lf" are used for using double and
long double using scanf/printf  Exponent can be from -127 to +126
 So the range is from 2-127 to 2126, or 10-38 to 10+38 approx.

9 10
Krithika Venkataramani ([email protected]) Krithika Venkataramani ([email protected])

Errors in representing real numbers Range of different data types

 There are three types of errors:  Variables are stored in a predefined space
 Underflow: Trying to store exponent less than -127  A unit of storage is a Byte
 A Byte has space to store a sequence of 8 binary digits
 Overflow: Trying to store exponent more than 126
 Different variable types have different storage space assigned
 Rounding off: Storing the nearest floating point number  Assignment of space is machine dependent
 Floating point arithmetic Type Space assigned in Bytes Range
 The hardware has to do a lot more for floating point arithmetic
char 1 - 27 to (27-1)
compared to integer arithmetic
unsigned char 1 0 to (28-1)
 Do not store numbers as floating point, unless you really
short int 2 - 215 to (215-1)
need fractions
unsigned short int 2 0 to (216-1)
int 4 - 231 to (231-1)
unsigned int 4 0 to (232-1)
float 4 (approx) [10-38, 1038]
double 8 (approx) [10-308, 10308]
11 12
Krithika Venkataramani ([email protected]) Krithika Venkataramani ([email protected])

3
 3/13/2012

Input and output of variables Character type

 Correct type specification must be used  Variable type ‘char’ used for representing characters
Type Format Specifier
 Characters are special integers of much shorter size
char %c
 Only 256 characters can be represented
int %d
 Digits 0-9 are not represented by 00000000 - 00001001
unsigned int %u
float %f,, %g,
g, %e  0-9 represented by a continuous sequence
double %lf  Similarly A-Z (a-z) also represented by a continuous
long double %Lf sequence
 scanf is for input  ASCII character set is most widely used
 Format: scanf(“<specification>”, &<name>);  specifies a standard that maps characters to numbers 0-127
 E.g. c is a char: scanf(“%c”, &c);  Extended ASCII assigns symbols to numbers 128-255
 printf is for output  ASCII and Extended ASCII use 1 Byte for storage
 Format: printf(“<specification>”, <name>);  Unicode includes characters from all languages of the world
 E.g. c is a char: printf(``%c'', c);  Unicode uses 2 Bytes
13 14
Krithika Venkataramani ([email protected]) Krithika Venkataramani ([email protected])

The ASCII Table Printing the Code of a Character

/*Program to print the code of a character*/

#include <stdio.h>
void main()
{
int code; //Declare variable to store the code
code = (int) getchar(); //Asking user to input the character
printf("%d", code); //printing the code of the character
}

15 16
Krithika Venkataramani ([email protected]) Krithika Venkataramani ([email protected])

4
 3/13/2012

Additional data types Additional formats for octal and hexadecimal

Data type Format specifier Size (machine Range
dependent)
int %d(decimal), %i 4 bytes -231 -1 to 231 -1
Data type Format Display/ Read
unsigned int %u 4 bytes 0 to 232 -1
specifier
short int (unsigned) %hd (%hu) 2 bytes -215 -1 to 215 -1
unsigned int %o unsigned octal integer
long int (unsigned) %ld (%lu) 8 bytes -263 -1 to 263 -1
char %c %d
%c, 1 byte -128 to 127 unsigned
g int %x, %X unsigned
g hexadecimal integer
g
unsigned char %u, %d 1 byte 0 to 255 unsigned long int %lo unsigned octal integer
string %s array of -- unsigned long int %lx, %lX unsigned hexadecimal integer
characters
unsigned short int %ho unsigned octal integer
float %f, %g, %e 4 bytes 3.410-38 to 3.41038
unsigned short int %hx, %hX unsigned hexadecimal integer
double %lf, %lg, %le 8 bytes 1.710-308 to 1.710308
long double %Lf, %Lg, %Le 16 bytes ?
1 bit: 1 or 0 1 Byte: 8 bits
Note: for 32-bit machines long int and int are same 17 18
Krithika Venkataramani ([email protected]) Krithika Venkataramani ([email protected])