Fundamental Data Types

1
Outline
1. Motivation
2. Integral (Integer) Types
3. Floating Point Numbers
4. Type Conversion
5. Back to the Original Example

2
1. Motivation
• After our previous lecture, you should be able to have intuitive understanding of
the following simple C program

1 #include <stdio.h>
2
3 int main(void)
4 {
5 int area, h, b;
6
7 h = 3;
8 b = 4;
9 printf("Height and base of triangle: %d, %d\n", h, b);
10 area = (1/2) * h * b;
11 printf("Area of triangle: %d\n", area);
12
13 return 0;
14 }
3
1. Motivation
• This C program, however, do not produce the expected result. Why?! All the
formula looks alright to you, right?

1 #include <stdio.h>
2
3 int main(void)
4 {
5 int area, h, b;
6
7 h = 3;
8 b = 4;
9 printf("Height and base of triangle: %d, %d\n", h, b);
10 area = (1/2) * h * b;
11 printf("Area of triangle: %d\n", area);
12
13 return 0;
14 }
4
1. Motivation
• To understand the problem of the above program, you need to
learn more about data types
• In the previous lecture, we have introduced the integer data type:

int x = 0;

• You may also recall a double data type:

double pi = 3.14;

• How do they differ? Why do we need to have different data types?

• Understanding data types thoroughly is a crucial first step of being a
programmer in C

5
2. Integral (Integer) Types
Key concepts
• The smallest/largest integer values of type int

• Variation of integer types

– Integer types of different sizes
– Unsigned and signed integers

• Integer overflow

6
2.1. Basic Building Block: Bits
• Ultimately, a computer represent data in bits.

• A bit (binary digit) is the smallest unit data.

• A bit can be either 0 or 1.

• 1 byte == 8 bits

• Computers use combination of bits to represent all sorts

of data.
7
2.1. From Bits to Integers
• Binary number system uses N bits to represent integers.
– Typically, N = 8, 16, 32, 64 (corresponding to 1, 2, 4 and 8 bytes)

• E.g. 8 bits represent one of 28=256 different bit patterns:

00000000, 00000001, 00000010, 00000011, …, 11111110, 11111111
Some people assign each a non-negative integer value, 0, 1, 2, …, 254, 255.

Some people rearrange the patterns in a different order:

10000000, 10000001, 10000010, 10000011, 10000100, 1000101, …,

11111110, 11111111, 00000000, 00000001, 00000010, …, 01111111

And call them -128, -127, -126, -125, …, -2, -1, 0, 1, 2,3, …, 125, 126, 127
The patterns are used for representing both –ve, 0 and +ve integers.
8
2.1. From Bits to Integers
• With N bits, we can represent 2N distinct values.
– Half for negative integers, and half for non-negative integers:
-(2N-1), -(2N-1 - 1), …, -2, -1, 0, +1, +2, …, +(2N-1 - 1)

• Type int is typically 32 bits in size nowadays. As such, it

can represent integers in the following range

-(231), -(231 - 1), … , -2, -1, 0, +1, +2, …, +(231 - 1)

or
-2147483648, …, -2, -1, 0, 1, 2, …, 2147483647

9
2.2. Variations of Integral Types
Type Size in bytes Range
[Replit (2021)]
signed char 1 -128 to 127 (a signed byte)
short (or short int) 2 -215 to 215-1 (-32768 to 32767)
int 4 -231 to 231-1 (if 4 bytes)
long (or long int) 8 -263 to 263-1 (if 8 bytes)
unsigned char 1 0 to 255 (an unsigned byte)
unsigned short 2 0 to 216-1 (0 to 65535)
unsigned int
4 0 to 232-1 (if 4 bytes)
(or unsigned)
unsigned long 8 0 to 264-1 (if 8 bytes)

Why are there so many different types of integers?

10
2.2. Variations of Integral Types
• What is the appropriate type to represent integers
in this program?
– When the amount of data to be processed is large and
the memory space is scarce, we have to be mean.

• For now it suffices to know that these variations of

integral types exist. For most applications, using
int is adequate.

11
2.3. Integer Overflow
• Integer overflow occurs when the result of an
arithmetic operation is too large to be represented by
the underlying integer representation.

• e.g.: assume integers are 32 bits in size

– Add one to the largest signed positive integer:

2147483647 + 1 → –2147483648

– The correct result +2147483648 is NOT representable in 32-

bit signed integer representation, i.e. out of range.
• We have to use another data type to represent such a number

12
3. Floating Point Numbers
Key Concepts
• Floating point numbers representation and arithmetic are
not exact.

• [Advanced and Optional] Further Reading and Reference

IEEE 754 Standard for Floating-Point Arithmetic, Wikipedia

(https://en.wikipedia.org/wiki/IEEE_754)

IEEE 754-2019 - IEEE Standard for Floating-Point Arithmetic

(https://ieeexplore.ieee.org/document/8766229)

13
3.1. Floating Point Number Representation
• Floating point numbers and integers have different
representations.

• Not all real numbers are representable.

– Finite number of bits vs. infinitely many real numbers
– Decimal to/from binary conversion error, e.g. convert 0.1 to binary

• Floating-point number representation and arithmetic

operations need not be exact.
e.g., printf("%.20f; %.20f", 3.3, 2.1 - 2.0 - 0.1);
yields 3.29999999999999982236; 0.00000000000000008327

• For very large computations, rounding errors may accumulate

and become significant.
14
3.2. C Language Floating-Point Types
Type Size Range and Precision
[Replit (2021)]
4 bytes Range: ±3.4 x 10± 38
float
32-bit single-precision Precision: 6 significant decimal places
8 bytes Range: ±1.7 x 10± 308
double
64-bit double-precision Precision: 15 significant decimal places
long System dependent; may provide wider
≥ 8 bytes [16 bytes]
double range and more precision

• 123.451234512345 is more precise than 123.451

• For most applications, using double is recommended.
• If possible, avoid using float which is imprecise in
modern day standard, thus leading to loss of precision.
15
3.3. Using floating point with printf()
• Remember our example last time?
1 #include <stdio.h>
2
3 int main(void)
4 { A) 3.141593
5 double pi = 3.1415927; B) 3.14
6 C) 3.1415927
7
8 printf("A) %f\n", pi);
9 printf("B) %.2f\n", pi);
10 printf("C) %.7f\n", pi);
The format specifier, %.xf, tells
11 printf() to format the
12 return 0; corresponding floating point
13 } number with x decimal places.
14

16
3.3. Using floating point with printf()
• You should now realize that integers and floating
point numbers are completely different things in C
language
• You must specify correctly what data type you are
supplying or you will see meaningless results
• What do you think the result will be in the
following example?
e.g. printf("%d", 3.14);

17
3.4. Using floating point with scanf()
1 #include <stdio.h>
2
3 int main(void)
4 { Variables r1 and r2 are of type double
5 double r1, r2;
6
7 printf("Enter two real numbers:\n");
8 scanf("%lf%lf", &r1, &r2);
Concerning double-typed values,
9
10 printf("r1 = %f\n", r1); scanf() uses %lf (‘ell’ f);
11 printf("r2 = %f\n", r2); printf() uses %f.
12 Look similar but slightly different!
13 return 0;
14 }

Enter two real numbers: printf(), by default, prints

123 456.125
r1 = 123.000000
decimal numbers (floating point
r2 = 456.125000 numbers) with 6 decimal places.18
4. Type Conversion
Key Concepts
• How types are converted in an expression with mixed
types of numbers
e.g., 2.5 + 5 / 2 = ?

• How a double type value is converted to an integral

type value

• Explicit Type Conversion (Type Casting)

19
4.1. Expressions with mixed types of data
• HK$1000 + US$100 =?

• 3.1 + 2 =?

• double d = 4; What value will d hold?

• int x = 4.1; What value will x hold?

• Some kind of conversion is needed to ensure the type of

both operands are compatible before the computer can
evaluate the expressions.

20
4.1. Implicit Type Conversion
• C language has a set of conversion rules to resolve
certain mismatched operand types.

• As a convenient to programmers, compilers

automatically convert the value of an operand from one
type to another based on these rules whenever possible.

• Sometimes called coercion.

21
4.1. Arithmetic Conversions (Simplified Rules)
• If either operand is a double, the other is converted to
double. The result type is also double.
e.g.:
3.1 + 2 (3.1 is of type double)
= 3.1 + 2.0 (Therefore, 2 is converted to 2.0)
= 5.1 (Result is of type double)

• If both operands are of one of the integral types char,

short and int, then both operands are converted to
int. The result type is also int.

22
4.2. Converting Integral Type to double
• Converting integral type to double is safe.
– No warning is given at compile time

double d;
d = 4;
/* 4 is converted to 4.0,
and then 4.0 is assigned to d */

23
4.2. Converting double to Integral Type
• Converting a double to an integral type may result in
loss of data.
– If the number is within the range of the integral type, the
fractional part is truncated, i.e. discarded.
– Compilers usually warn at compile time (but NOT guaranteed.)

int x = 4.1; /* 4.1 is converted to 4 */

x = 2 * 4.1; /* 8.2 is converted to 8 */
x = 92345678901.2; /* behaviour is undefined
when the value is too
large to hold in x */

24
4.3. Explicit Type Conversion (Casting)
Syntax: (new_type) operand

• Converts the value of operand to the equivalent value

of type new_type.
– (new_type) is called the type casting operator
– Note that not every type conversion is possible, however.
e.g.,
double d = 4.2;
int y = (int) d; // y becomes 4, no warning
int x = d; // x becomes 4, compiler warns
// because of missing type casting

25
4.4. Type Conversion (Examples)
1 int x = 5, y = 2;
2 double a, b;
3 a = 2.5 + (x / y); /* R.H.S. is evaluated as
4 2.5 + (5 / 2)
5 => 2.5 + 2
6 => 2.5 + 2.0
7 => 4.5
8 */
9
10 b = 2.5 + ((double)x / y); /* R.H.S. is evaluated as
11 2.5 + (5.0 / 2)
12 => 2.5 + (5.0 / 2.0)
13 => 2.5 + 2.5
14 => 5.0
15 */

Example 4.1. Expression with mixed data types

26
4.4. Type Conversion (Examples)
1 int x, y;
2 double a = 2.6, b = 2.4;
3 x = (int)(a + 0.5); // x is assigned 3
4 y = (int)(b + 0.5); // y is assigned 2
Example 4.2. Rounding floating point numbers to nearest integer

27
4.4. Using Type Casting Operators (Exercise)
• Average of N integers
// Consider the following declaration
int total, N;
double avg;
…
// Suppose we have obtained the value of N and
// calculated the total of N integers.
// Which of these will correctly calculate the average?
avg = total / N; // A
avg = (double)total / N; // B
avg = total / (double)N; // C
avg = (double)(total / N); // D
avg = (double)total / (double)N; // E

28
4.5. How are numbers converted?
(Apply to both implicit and explicit conversions)

• double to integral types

– Only retain the integer part (no rounding)
– e.g.:
int x = (int) 4.9; // x gets 4
int y = (int) -3.99; // y gets -3

• "Larger" integral types to "smaller" integral types

– Retain only the least significant bits (LS-bit)
– e.g.: 32-bit integer (int) to 16-bit integer (short)
00000000000000100000000000000011 = 13107510
0000000000000011 = 310
short x = (short)131075; // x becomes 3

29
5. Back to the Original Example
• Given your understanding of data types now, what is wrong with our original
example?

1 #include <stdio.h>
2
3 int main(void)
4 {
5 int area, h, b;
6
7 h = 3;
8 b = 4;
9 printf("Height and base of triangle: %d, %d\n", h, b);
10 area = (1/2) * h * b;
11 printf("Area of triangle: %d\n", area);
12
13 return 0;
14 }
30
5. Back to the Original Example
• Given your understanding of data types now, what is wrong with our original
example?

1 #include <stdio.h>
2
3 int main(void)
4 { Are we using the proper
5 int area, h, b; type? Are these variables
6 always integers?
How about this? What are
7 h = 3; we telling printf?
8 b = 4;
9 printf("Height and base of triangle: %d, %d\n", h, b);
10 area = (1/2) * h * b;
11 printf("Area of triangle: %d\n", area);
12 Does 1/2 give integer as result,
13 return 0; or floating point? Should we
14 } use, say, 0.5 instead?
31
Summary
• All number types have an intended purpose,
precision, and range.
– Choose a proper data type to represent data
– Beware of and prevent overflow

• Floating-point representation and arithmetic may not

be exact.

• Expressions with mixed types of data

– Automatic and explicit type conversion
– Number conversion (double to int)
32
 Appendix: Finding out the size of an integer
1 #include <stdio.h>
2
3 int main(void) {
4 printf("size of int = %d bytes\n", (int) sizeof(int) );
5 return 0;
6 }

size of int = 4 bytes

• sizeof(data_type) yields the number of bytes used

to represent a value of type data_type (as unsigned
long).
• (int) explicitly converts the value to int.

33
Reading Assignment
• C: How to Program, 8th ed, Deitel and Deitel
• Appendix C Number Systems

34
Reminder: PreLabs are Ready!
• Every Mon afternoon we will release the PreLabs
– Meant to help you prepare for the lab
– Due Wed 9:30am – Please try it after the lecture and
submit before Wed!
– Don't worry – it's super easy (takes < 30 min) and it's
very easy marks to get! Don't forget!

PreLabs are marked

"(PreLab)" on repl.it

35