float Data Type

1
float Data Type

• Data type that can hold numbers with decimal
  values
• e.g. 3.14, 98.6
• Floats can be used to represent many values
• Money (but see warning below)
• distance
• weight, etc.

2
Float Example (Program 2.2)

• / Float Example Program /
• include ltstdio.hgt
• int main ()
• float var1, var2, var3, sum
• var1 87.25
• var2 92.50
• var3 96.75
• sum var1 var2 var3
• printf ("Sum .2f", sum)

f indicates floating values .2f displays a
floating point value with 2 decimal
points. Output Sum 276.50
3
Example Finding an Average

• Suppose you want to determine a students
  average.
• int num 4
• float avg 909295100/num
• Problem 1 Operator Precedence
• By rules of operator precedence, 100/4 is
  evaluated first. Hence, avg is set to 302.
• To solve this problem, use ()
• float avg (909295100)/num

4
Finding an Average

• Problem 2
• 90, 92, 95, 100 and 4 are all integers. Hence,
  this is integer division.
• Integer division can result in data truncation
  (or loss of data.)
• Hence, avg is set to 94.0, but the students
  real average is 94.25.
• There are actually three ways to solve this
  problem.

5
Rules of Promotion

• Promotion when mixing ints and floats,
  everything is promoted to floats.
• In our average example, there are three ways to
  force promotion, and get the right answer of
  94.25
• 1. change num to a float
• float num 4.0
• float avg (909295100)/num

6
Rules of Promotion

• 2. Use a Cast Operator
• int num 4
• float avg (909295100)/(float)num
• In this case, num is explicitly cast to a float.
  And, because we have one float, everything else
  is promoted. Note that you can also use the
  (int) cast to cast a float to an integer.
• 3. Divide by 4.0
• float avg (909295100) / 4.0

7
Finding an Average

• Problem 2
• 90, 92, 95, 100 and 4 are all integers. Hence,
  this is integer division.
• Integer division can result in data truncation
  (or loss of data.)
• Hence, avg is set to 94.0, but the students
  real average is 94.25.
• There are actually three ways to solve this
  problem.

8
NEVER USE floats HERE

• In a condition for equality comparison
• Floats may be represented differently from what
  you think by the computer
• E.g. 1.9 to you may be 1.89999999999999999999999
• 1.9 will not necessarily equal 1.9!
• In critical calculations for the same reason
• E.g. .1 added 10 times often will not add up to 1
• Use long ints instead and keep track of where
  your decimal point is (e.g. 1.75 should be
  stored as 175)

9
Comparing floats

• If you need to compare floats, determine some
  threshold and test against that
• E.g.
• if ( (fDollarsReceived fDollarsExpected) lt
• .00001) )
• printf( OK, close enough to equal
• for me!\n )
• Note you have to check for the reverse too

10
Comparing floats contd

• Complete working examplefloat fDollarDiff 0
• fDollarDiff fDollarsReceived
• fDollarsExpected
• if ( -.00001 lt fDollarDiff
• fDollarDiff lt .00001 )
• printf( OK, close enough to equal
• for me!\n )

11
Good case for a constant

• By the way that .00001 is a prime candidate for
  define!
• E.g
• define f_ACCEPTABLE_THRESHOLD .00001

12
printf conversion specifications

• The format control string (the first argument to
  printf) contains
• conversion specifiers (d, i, f)
• field widths
• precisions
• other info we will not look at

13
Field width

• The exact size of the field in which data is to
  be printed is specified by a field width.
• If the field width is larger than the data being
  printed, the data will normally be right-
  justified within that field.
• Use a - sign before the number in field width
  to left-justify
• If the field width is smaller than the data being
  printed, the field width is increased to print
  the data.

14
Field width contd

• An integer representing the field width is
  inserted between the percent sign () and the
  conversion specifier
• For example, 4d
• int iYourAnswer 1
• printf( The answer 4d is correct!\n,
• iYourAnswer )
• Will print the following
• The answer 1 is correct!

15
Precision

• You can also specify the precision with which
  data is to be printed.
• Precision has different meaning -- for different
  types.
• integer minimum number of digits to be printed
• float number of digits to appear after the
  decimal point

16
Precision contd

• To use precision, place a decimal point (.)
  followed by an integer representing the precision
  between the percent sign() and the conversion
  specifier (ie .2f)

17
switch Multiple-Selection Structure

• Used when testing a variable or expression for
  EQUALITY (ie no gt, lt, gt, lt tests) separately
  for each of the constant integral values it may
  assume.
• Preferred over if else in situations where you
  are testing the same expressions for equality
  with many different values.
• Allows you to perform different actions for each
  test.

18
switch Multiple-Selection Structure

• switch (expression)
• case value1
• action(s)
• break
• case value2
• action(s)
• break
• case default
• actions(s)
• break

keyword switch
expression can be a variable or a more
complicated expression
could use more than one case if the same actions
are required
actions within a single case do not need brackets
the default case will be executed in the event
that no other case is
19
beware of fall through

• If you forget to use the break keyword between
  cases, unexpected things may happen.
• Once a case tests true, all the statements
  following that case, will be executed until the
  next break.
• Experienced programmers may use this on purpose.
  For this class we will not.