Datatypes in C

ESC101
9th August
 Last Class
• Basic alphabet of C
– Character set
– ASCII encoding
• Constructing words from alphabets
– Identifiers
– Keywords
 This Class
• C data types
• Type conversions
– Implicit
– Explicit
• Consequences
• Related reading: Chapter 2.7, 2.8, 2.13 in
Gottfried
 Data Types in C
• char
– Single character, e.g. a or C or 6 or $
• int
– Bounded integers, e.g. 732 or -5
• float
– Real numbers, e.g. 3.14 or 2.0
• double
– Real numbers with more precision

9/19/2019 4 Esc101, Programming

 Type Modifiers in C
• Signed
– Range [-2N-1 , 2N-1-1]
• Unsigned
– Range [0,2N-1]
• Short
– Can use half the size of the normal data type
• Long
– Can use double the size of the normal data type
 Composite data types
• signed short int = signed short = short (%hi)
• signed long int = signed long = long (%li)
• unsigned int (%u)
• float (%f)
• double (%f or %lf)
• long double (%Lf)
 Floating Point Representation
• Have to represent three things
– sign
– Number
– Exponent
• Assign some bits of memory for each
– 1 bit for sign
– m for exponent
– n for mantissa
 Conceptual Example
• Consider a 4 bit memory
– What can you assign with unsigned int?
• 0,1,.....15
– What can you assign with signed int?
• Use twos complement notation
• -8,-7,.... ,7
– What can you assign with float?
s e m m 1.0, 1.1, 1.2, 1.3
2.0, 2.2, 2.4, 2.6
-1.0, -1.1, -1.2, -1.3
(-1)s * 1.m* 2e-0 -2.0, -2.2, -2.4, -2.6
 Edge Cases in float Representations
• Has upper bound
• Has lower bound
• Needs special handling for special numbers
– Zero (when e and m are all zeros)
– Infinity
• Exact matches can be problematic
– Is x = 0.902323?
 temp_conversion.c • Microprocessors represent real
numbers using finite precision, i.e.,
Compile and Run using limited number of digits after
decimal point.
# include <stdio.h>
int main() { • Typically uses scientific notation:
float C; 12.3456789 represented as
float F; 1.23456789E+1.
C=50;
F = ((9*C)/5) + 32;
“%f” signifies that the corresponding
printf(“The temperature”);
variable is to be printed as a real number in
printf( “ %f ”, C);
decimal notation.
printf(“Celsius equals”);
printf(“ %f ”, F);
printf(“Fahrenheit”); C 50.000 F
return 0; ?? ??
122.0000
}

The temperature 50.000000 Celsius equals 122.000000 Fahrenheit

IEEE 754 Floating Point Representation
 Single-precision Floating Point
Representation

This is what you’re using when you are invoking float

 Double Precision
• Same logic as single precision, but with 64 bits
of memory
 Type Conversion (Type Casting)
• Converting values of one type to other.
– Example: int to float and float to int
(also applies to other types)
• Can be implicit or explicit
int k =5;
float x = k; // good implicit conversion, x gets 5.0

float y = k/10; // poor implicit conversion

float z = ((float) k)/10; // Explicit conversion, z gets 0.5

 Typecasting Application
int main() { int main() {
int total=100, number=50; int total=100, number=50;
float percentage=0.0; float percentage=0.0;
percentage=(number/total)*100; percentage= (float) number/total*100;
printf("%.2f",percentage); printf("%.2f",percentage);
return 0; return 0;
} }

Output: 0.00 Output: 50.00

 Loss of Information!

• Type conversion may result in lost information

• Larger sized type (e.g. float ) converted to

smaller sized type (e.g. int) is undefined/
unpredictable

9/19/2019 16 Esc101, Programming

 float to int: type conversion (result ok)
float x;
#include<stdio.h>
…
int main() {
(int) x;
float x; int y; /* define two variables */
x = 5.67;
converts the
y = (int) x; /* convert float to int */
real value
printf(“%d”, y);
stored in x into
return 0;
an integer. Can
}
be used
anywhere an
int can.
Output : 5
 float to int Type Conversion (not ok!)
• float is a larger box, int is a The floating
smaller box. Assigning a float to point number
1E50 is too
an int may lead to loss of
large to fit in
information and unexpected an integer box.
values.
# include <stdio.h>
int main() { Output:
float x; int y; 2147483647 Careful when
x = 1.0E50; // 1050 converting from a
y = (int) x; ‘larger’ type to
printf(“%d”, y); ‘smaller’ type.
return 0; Undefined.
Rocket Science
• First launch of the Ariane
5 rocket on 4th June 1996
• Rocket lost its flight path
and disintegrated 40
seconds into launch
• Cost  $370 million
• Fundamental cause of
disaster
– Float to int data type
casting
Rocket science
• Rocket’s horizontal velocity
was greater than older rocket’s
(Ariane 4)
• Inertial navigation system
– A single calculation required
conversion from double to int
– Worked fine with the smaller
value in Ariane 4
– Gave a garbage result for the
large value in Ariane 5
• Commanded engine to make
an impossibly large course
correction
• Veered off course
– Auto-destruct sequence
initialized
 char to int
• Range: 0 to 255
• You should NOT try to remember ASCII
values
– Encoding/programming languages provide
alternatives to use them
• C treats characters as integers corresponding
to their ASCII value
• While displaying with %c placeholder, the
ASCII value is converted to its corresponding
character
9/19/2019 21 Esc101, Programming
 ASCII character set

Translates letters to numbers for the computer to

understand
 char  int

• conversion between character and integer

datatypes can be exploited to write programs.

Output:
printf("%d\n", 'A'); 65
55
printf("%d\n", '7'); F
printf("%c\n", 70); 321 is outside range! What
printf("%c\n", 321); do you think will be the
output of
printf("%c\n",321);
Try it out
9/19/2019 23 Esc101, Programming
 char-int operations

• Interconversion between character and

integer datatypes can be exploited to write
programs.
Output:
printf("%c\n", ‘C’+5); H
printf("%c\n", ‘D’ - ‘A’ + ‘a’ ); d
53
printf("%d\n", ‘3’ + 2);

• Placeholder determines the output.

• Use with caution.
• Avoid arithmetic operation such as * and / on characters.
•9/19/2019
Common Mistake: Incorrect data
24
type of placeholder.
Esc101, Programming
 Next Class
• C program structure
– Variables
– Declarations
– Expressions
– Statements
• Related reading: Chapter 2.9-2.12 in Gottfried