# Remove all variables.
# It's a good idea to start with a "clean slate".
rm(list=ls()) # remove all variables
####################################################################.
####################################################################.
#
# INTRO TO USER DEFINED FUNCTIONS ####
#
####################################################################.
####################################################################.
#----------------------------------------------------------------------------.
# R has many, many built-in functions. However, if you don't find what you
# are looking for with R's built-in functions, you can also define
# your own functions!
#
# The following simple examples demonstrate the concepts.
#
# EXAMPLE:
#
# The following is called a FUNCTION DEFINITION. ####
# It defines a brand new function called "doubleIt".
# The doubleIt function is defined to take a single argument named nums.
# The doubleIt function returns two times whatever value is passed to nums.
# (see examples below)
#
# doubleIt <- function(nums) {
# return(2 * nums)
# }
#
#----------------------------------------------------------------------------.5 5. User defined functions
5.1 Type function definitions in the SCRIPT window
# YOU SHOULD TYPE FUNCTION DEFINITIONS IN THE SCRIPT WINDOW ####
# (NOT IN THE CONSOLE WINDOW)
#----------------------------------------------------------------------------.
# Since function definitions tend to be rather long and span multiple lines,
# it's not very convenient to type them in the Console window.
# Rather type them in RStudio's script window and then run the code
# To run the code, put your cursor on the first
# line of the function definition in RStudio's "script window"
# (or highlight the complete function definition) and then
# press ctrl-ENTER (windows) or cmd-ENTER (Mac). After you run the
# code, you will see the name of the function
# appear in RStudios "Environment" window. You can also see the name of the
# function in the output of the ls() command.
#
# To get R to recognize that this new function exists,
# you MUST run the function definition - i.e. put your cursor on the first
# line of the function definition (or highlight the complete function definition)
# in the script window and then press ctrl-ENTER (windows) or cmd-ENTER (Mac).
#
# It's very important to realize that it's not enough to just type the code
# in the script window. Don't forget to run the code for the function definition
# (many newbies forget this part).
#
# When you run the code for the function definition, all that appears to happen is
# that you will see the code displayed in the console window. Nothing else
# appears to have happened. For example:
#
# > doubleIt <- function(nums) {
# + return(2 * nums)
# + }
#
# The "+" signs at the beginning of the line are NOT part of the code.
# They simply mean that your function definition spans multiple lines of code.
#---------------------------------------------------------------------------5.2 Run the definition before using the function
#----------------------------------------------------------------------------.
# USING YOUR NEW FUNCTION
#----------------------------------------------------------------------------.
# Once you've run the code for the function definition (see above)
# you can then use the new function with any values that you wish. For example:
#
# > doubleIt(7)
# [1] 14
#
# > doubleIt( c(10,20,30) )
# [1] 20 40 60
#
# > doubleIt( c( 1:3 , rep(5,6) ) )
# [1] 2 4 6 10 10 10 10 10 10
#----------------------------------------------------------------------------.
# define a new function
#
# DON'T FORGET TO RUN THE CODE FOR THE FUNCTION DEFINITION ####
# i.e. put your cursor on the first line and
# press ctrl-ENTER (win) or cmd-ENTER (mac)
doubleIt <- function(nums) {
return(2 * nums)
}
# We can now use the doubleIt function with many different values similar
# to the way you can use R's built-in functions.
#
# Before running the following lines of code you must make sure that you've
# run the function definition above.
doubleIt(3) # 6[1] 6doubleIt(c(10,20,30)) # 20 40 60[1] 20 40 60doubleIt( nums = c(100,500)) # you can specify the name of the argument in the function call[1] 200 1000doubleIt(c(100,500)) # same thing[1] 200 1000# In general you can now work with the user-defined-functions
# the same way you would work with built in functions.
aBigNumber = doubleIt(1000) + 210
aBigNumber[1] 2210# As with all functions, you must figure out the value of the arguments
# before the function is actually called.
doubleIt( seq(1,3) )[1] 2 4 6# original: doubleIt( seq(1,3) )
# seq: doubleIt( c(1,2,3) )
# doubleIt: 2 4 6
doubleIt( 1 + seq(1,3) )[1] 4 6 8# original: doubleIt( 1 + seq(1,3) )
# seq: doubleIt( 1 + c(1,2,3) )
# addition: doubleIt( c(2,3,4) )
# doubleIt: 4 6 8
doubleIt( c( 1:3 , rep(5,6) ) )[1] 2 4 6 10 10 10 10 10 10# original: doubleIt( c( 1:3 , rep(5,6) ) )
# expand 1:3: doubleIt( c( c(1,2,3) , rep(5,6) )
# rep function: doubleIt( c( c(1,2,3) , c(5,5,5,5,5,5) ) )
# outermost c function: doubleIt( c(1,2,3,5,5,5,5,5,5) )
# doubleIt: 2 4 6 10 10 10 10 10 105.3 “Local” variables
##################################################################.
#
# The code inside a function definition may consist of several lines
#
# You may use "local variables" in a function definition.
#
##################################################################.
# Below we defined the function, trippleIt, that tripples the values
# in the argument, nums.
#
# We wrote trippleIt using a slightly different approach than we used
# for doubleIt. Instead of simply multiplying by 3, we
# used a variable named total to store the
# value that will be returned. Every line of the function
# adds another multiple of the values in nums to the total.
# The last line of the function returns the total.
#
# This seems unnecessarily complex. It is. We could have very easily written
# trippleIt the same way that we wrote doubleIt, i.e. by simply
# multiplying by 3. That would work just fine.
# We used this alternative approach simply to demonstrate some concepts:
#
# (a) a function definition can contain multiple lines of code
#
# (b) You may define "local variables" inside a function definition.
# A "local variable" is a variable that is created inside a
# function definition to help perform the calculations for
# the function. We'll discuss more about local variables later below.
#
# Define the function
trippleIt <- function(nums) {
return(3 * nums)
}
# call the function
trippleIt(c(5,6,7))[1] 15 18 21# A differnt version of trippleIt that returns the same values as the
# original version above. The purpose of writing the following version of
# trippleIt is simply to demonstrate that a function can have multiple lines
# of code.
#
# This will also be helpful to demonstrate R's "debugger" (which we will
# see soon)
trippleIt = function( nums ){
total = nums
total = total + nums
total = total + nums
return (total)
}
trippleIt(3) # 9[1] 9trippleIt(c(10,20,30)) # 30 60 90[1] 30 60 90# to use the debugger for a function, you need to call the debugonce function
# to tell the debugger to debug the specified function the next time it is called.
trippleIt(c(10,20,30))[1] 30 60 90# You can use the trippleIt and doubleIt functions together.
trippleIt(doubleIt(3))[1] 182 ^ doubleIt(c(1, 1.5, 2))[1] 4 8 16trippleIt( 2 ^ doubleIt(c(1, 1.5, 2)) )[1] 12 24 48#----------------------------------------------------------------.
# "LOCAL VARIABLES" AND "SCOPE OF VARIABLES" ####
#----------------------------------------------------------------.
# Note that variables that are created inside the definition of
# a user-defined-function are only "known" inside of that function
# definition. When the trippleIt function above finishes executing, the variable
# total does NOT appear in the variables in the environment.
# For example ls() will not show the total variable.
#
# A variable that can only be used inside of a particular function definition ====
# is known as a "local variable" for that function. We say that the ====
# variable has "local scope". ====
#
# We will discuss much more about local variables and
# the "scope" of a variable a little later (below).
#----------------------------------------------------------------.
ls() # show all variables - note that total isn't in the list.[1] "aBigNumber" "doubleIt" "trippleIt" # We will return to discuss these issues more in depth a little later.5.4 Last value of the function is returned
#######################################################################.
# If you don't write the word "return" then
# the final value of the function is returned as the value of the function. ####
#######################################################################.
# We could have written the definition of trippleIt without the return.
# By default, if there is no return statement then
# the last value of the function is what is returned.
# The following is an equivalent definition of the trippleIt function.
# Even though we left out the "return" the last value is still returned
# as the value of the function.
trippleIt = function( nums ){
total = nums
total = total + nums
total = total + nums
total # we took out the word return, it still works
}
# This still works
trippleIt(3) # 9[1] 9trippleIt(c(10,20,30)) # 20 40 60[1] 30 60 905.5 Don’t return an assignment statement
# BE CAREFUL - don't return the value of an assignment statement ####
#
# If we take out the last line of the trippleIt function entirely
# (e.g. by commenting it out) then the last command of the function
# becomes an assignment statement.
#
# However, remember that when you run an assignment statement no value
# is actually displayed to the screen. For example, if you run
#
# > x = 5
# >
#
# Nothing is displayed as the result of the command. In a similar way, if you
# final command of the function is an assignment statement, nothing
# will be displayed when you run the function.
trippleIt_returningAnAssignment = function( nums ){
total = nums
total = total + nums
total = total + nums
# total # if we remove this entire line, the function does not work the same way
}
# Now since the last command to be executed in trippleIt is an assignment
# statement, the following lines do NOT display any values.
trippleIt_returningAnAssignment (3) # nothing is displayed
trippleIt_returningAnAssignment (c(10,20,30)) # nothing is displayed
# As a general rule, you should return the value of an assignment statement
# or have an assignment statement as the last value of a function
# whose job it is to return a value.
###########################################################################.
# 2023 - WILF - UP TO HERE - AFTER CLASS 5 ####
###########################################################################.
###########################################################################.
# 2023 - BEREN - UP TO HERE - AFTER CLASS 6 ####
###########################################################################.5.6 “invisible” return values
#--------------------------------------------------------------------------.
# Note about "invisible" return values ####
#--------------------------------------------------------------------------.
# What is very strange is that even though in it's current form, the
# trippleIt function does NOT display anything when you call it, if you
# assign the result of the function to a variable, the variable WILL
# contain the value!!! This may be very surprising.
trippleIt_returningAnAssignment (3) # nothing is displayed
x = trippleIt_returningAnAssignment (3)
x # 9[1] 9# The "value" of an assignment statement is an "invisible" value. ####
# ("invisible" is an little known R concept that many people don't understand).
# Invisible values can be captured to a variable, but by default they are not displayed. ####
#
# This is a somewhat advanced topic so we will not discuss it more now.
# However, you should avoid ending a function with an assignment
# statement if the purpose of the function is to return a value
# that needs to be displayed.5.7 don’t pass too many arguments
#------------------------------------------------------------------.
# Be careful - don't pass too many arguments ####
#
# The following will generate an error since you are trying to pass
# three different values to trippleIt but trippleIt is only
# defined to take a single argument, nums.
#------------------------------------------------------------------.
trippleIt( 100, 200, 300 ) # ERROR - trippleIt only allows for one argument, numsError in trippleIt(100, 200, 300): unused arguments (200, 300)# This fixes the problem since c(100,200,300) is a single vector.
# The vector c(100,200,300) is passed to nums.
trippleIt( c( 100, 200, 300) ) # 200 400 600[1] 300 600 9005.8 Function name without parentheses() shows definition code
##############################################################################.
# A function's name without any parentheses, shows the definition of the function. ####
##############################################################################.
doubleIt # show the definition of doubleItfunction(nums) {
return(2 * nums)
}
<bytecode: 0x0000020a2ff0ed50>5.9 — Practice —
##################################################.
# QUESTION - part (a) ####
#
# A professor gives 3 tests.
# The first test is worth 25% of the grade,
# the 2nd test is worth 30% of the grade and the third test is worth 45% of the grade.
# A student in the class got the following test grades are 80, 90 and 100.
# Use the variables, weights and grades, shown below to calculate the student's
# overall "weighted average".
##################################################.
weights = c(0.25, 0.3, 0.45)
tests = c(80,90,100)
# ANSWER
sum(weights * tests) # ANSWER[1] 92##################################################.
# QUESTION - part (b) ####
#
# Write a function named, weightedAvg, that takes
# two arguments, weights and tests. The function
# should work as shown below.
#
# > WRITE THE FUNCTION DEFINITION HERE
#
# # Now you can use the function
# > weightedAvg( c(0.25, 0.3, 0.45) , c(80,90,100) ) # result should be 92
# [1] 92
#
# > weightedAvg( c(0.25, 0.3, 0.45) , c(70,85,50) ) # result should be 65.5
# [1] 65.5
#
# i. Write the function with the return statement
#
# ii. Write the function without the return statement
# (should produce the same results)
##################################################.
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.
# step 1 - define the function
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.
weightedAvg = function (grades, weights){ # ANSWER
return( sum(grades * weights) ) # ANSWER
}
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.
# step 2 - you must RUN the function definition.
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.
# As with all code that you write in the script window, to run it
# simply, put your cursor on the first line of the function definition and
# press ctrl-ENTER (win) or cmd-ENTER (mac).
#
# After you run the function definition you should see the
# name of the function appear in the Environment tab (by default
# it's found in the upper right hand corner of the RStudio window)
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.
# step 3 - use the new function to perform some calculations
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.
# Call the function by just supplying the values for
# the tests and the weights. You do not have to type the argument names
# as long as you supply the values in the proper order.
weightedAvg( c(80,90,100), c(0.25, 0.3, 0.45)) # ANSWER[1] 92# Just as with the built in functions you may also supply the
# argument names when you call the function. This makes the code
# more clear but it is NOT necessary as long as you supply
# the values in the correct order.
weightedAvg( grades=c(80,90,100), weights=c(0.25, 0.3, 0.45)) # ANSWER[1] 92# You can also pass the values of the arguments as variables.
# Note that the variable names do NOT have to be the same as the
# argument names. The variable names simply represent the values that
# were assigned to them. For example:
t = c(80,90,100) # these are the test grades
w = c(0.25, 0.3, 0.45) # these are the weights
weightedAvg(t, w) # this works[1] 92weightedAvg(grades=t, weights=w) # this also works[1] 92# Just as with built-in functions, if you type the names of the arguments,
# you can write the arguments in a different order.
weightedAvg(weights=w, grades=t) # this also works[1] 92#--------------------------------------------------------------.
# ANSWER - Another way - without the return function
#---------------------------------------------------------------.
# The answer above used the return() function to "return" the final value
# from the function. We'll see later that sometimes it is necessary to
# use the return() function. However, with simple functions (as these are)
# by default the R function will return whatever value is the last value
# to be executed inside of the function.
weightedAvg = function (grades, weights){ # ANSWER
# In this version there is no "return" function call
# Since the following is the last line this is automatically returned.
sum(grades * weights) # same as return(sum(grades * weights))
}
# This still works
weightedAvg( c(80,90,100), c(0.25, 0.3, 0.45))[1] 925.10 Default argument values
###########################################################################.
#
# Default argument values. ####
#
###########################################################################.
# If the function definition specifies a value for an argument then that
# value will be used as the default argument if the call to the function
# doesn't specify that argument.
# For example, the following function returns the volume of a cylinder
# (e.g. a soda can) that has a specific height and a specific radius.
#
# The "defualt value" for the height is 1 and the default value of the
# radius is specified as 0.5 (i.e. a diameter of 1).
volumeOfCylinder = function( radius=0.5, height=1){
return( pi * radius^2 * height )
}
volume2(1,5) # volume of a cylinder with radius = 1 and height = 5Error in volume2(1, 5): could not find function "volume2"volumeOfCylinder(1,5) # volume of a cylinder with radius = 1 and height = 5[1] 15.70796volumeOfCylinder(2) # volume of a cylinder with radius = 2 and height = 1[1] 12.56637volumeOfCylinder() # volume of a cylinder with radius = 0.5 and height = 1[1] 0.7853982volumeOfCylinder(0.5, 1) # same thing[1] 0.7853982volumeOfCylinder(radius=0.5, height=1) # same thing[1] 0.7853982volumeOfCylinder( height = 6) # radius is 0.5 height is 6[1] 4.712389# Note that you can refer to the other arguemnts when defining the
# default values. This is shown in the following QUESTION/ANSWER.5.11 — Practice —
##################################################.
# QUESTION - part (c) ####
#
# Create the weightedAvg function described in the previous question
# so that if a value for weights is not supplied in the function call
# then equal weight should be given to each grade.
#
# Call the function weightedAvg_newVersion so that it doesn't
# replace the original weightedAvg function
##################################################.
# ANSWER
#
# Notice the default value of weights argument that appears on the first line.
# It is set so that the weight is evenly distributed among the grades.
weightedAvg_newVersion = function (grades, weights=1/length(grades)){ # ANSWER
return( sum(grades * weights) ) # ANSWER
} # ANSWER
# Using the new function - with the default weights
weightedAvg_newVersion(c(80,90,100)) # ANSWER[1] 90# Using the new function by specifying our own weights
weightedAvg_newVersion(c(80,90,100), c(0.25, 0.25, 0.5)) # ANSWER[1] 92.5weightedAvg_newVersion(c(80,90,100), 1/3) # ANSWER[1] 90# The original function required that you specify all the arguments
# The following call to the original function will generate the following errror:
#
# Error in weightedAvg(c(80, 90, 100)) :
# argument "weights" is missing, with no default
weightedAvg(c(80,90,100)) # ERROR - missing argumentError in weightedAvg(c(80, 90, 100)): argument "weights" is missing, with no default5.12 2023 - WILF - UP TO HERE - AFTER CLASS 6 (also covered debugger)
5.13 2023 - BEREN - UP TO HERE - AFTER CLASS 7 (also covered debugger)
5.14 Calculations in middle of function definition are NOT displayed
##########################################################################.
# Code in a function definition is DIFFERENT than code that is run
# directly in the console window:
#
# If you type a simple calculation IN THE CONSOLE the result is DISPLAYED.
# For example:
# > 10 + rep(1,3)
# [1] 11 11 11
#
# If you type a simple calculation IN A FUNCTION DEFINITION the result is IGNORED. ####
# See the examples below.
##########################################################################.
# A calculation typed in the console that would display a value has
# absolutely no effect on the code of a function.
f = function( stuff ){
3*10 # this line is ignored (it is not displayed)
5+7 # this line is ignored (it is not displayed)
stuff + 5 # this line is ignored (it is not displayed)
2 # the function always returns 2
}
f(10) # 2[1] 2f(c(100,200,300)) # 2[1] 2f(6547654.4356543) # 2[1] 2# You can save the value of the calculation in a variable and then
# continue to use the value by referring to the variable.
# e.g. see the following code (including the comments).
calculateBill = function( charge, tip ){
tax = 0.0875 * charge # you can save this calculation in a variable
return (charge + tax + tip) # now you can use the tax variable in this line
}
calculateBill(100, 10)[1] 118.755.15 — Practice —
##################################################.
# QUESTION
#
# Write a function named, gpa, to calculate a GPA.
# Feel free to use "local" variables to make your function
# easier to read.
#
# ARGUMENTS:
# credits - a vector that contains the number of credits for each course
# grades - a vector that contains the grades for each course (0 for F, 1 for D, 2 for C, 3 for B, 4 for A)
#
# e.g.
# > getGpa (credits = c(4,3,3), grades=c(2, 3, 4)) # i.e. (4*2 + 3*3 + 3*4) / (4+3+3)
# [1] 2.9
#
#
# *****************************************************************
# ** READ THE FOLLOWING IF YOU DON'T KNOW HOW TO CALCULATE A GPA **
# *****************************************************************
#
# A GPA is a weighted average.
#
# Letter grades are treated as numbers, i.e. A is 4, B is 3, C is 2, D is 1, F is 0
#
# To calculate a GPA you determine what percent of your overall
# college average is attributed to a specific class, then you multiply
# the grade for that class by that number, i.e.
# (credits_for_a_class / total_credits_taken) * grade_for_the_class
# Finally, you add up the results for all of the classes.
#
# EXAMPLE: physics 4 credits, grade is B
# data analytics 3 credits, grade is A
# gym 0.5 credits, grade is B
#
# total creidts is 4 + 3 + 0.5
# percent of total that is attributed to physics is 4 / 7.5
# Portion of GPA attributed to physics is (4/7.5) * 3 # remember B is 3
#
# Do this for all classes to get that the GPA is 3.4 calculated as follows
#
# (4/7.5)*3 + (3/7.5)*4 + (0.5/7.5)*3 which is 3.4
#
##################################################.
# ANSWER
# (You need to do this)
# AFTER YOU RUN YOUR CODE, CHECK YOUR ANSWER:
gpa( credits=c(4,3,0.5) , grades=c(3,4,3)) # 3.4 (if you wrote the code correctly)Error in gpa(credits = c(4, 3, 0.5), grades = c(3, 4, 3)): could not find function "gpa"gpa( c(4,3,0.5) , c(3,4,3)) # 3.4 (if you wrote the code correctly)Error in gpa(c(4, 3, 0.5), c(3, 4, 3)): could not find function "gpa"#######################################################################.
# QUESTION - use the gpa function that you defined above to
# calculate Joe's GPA. Joe took 4 classes. The credits and grades were as follows
# 4 credit class - got a C
# 3 credit class - got an A
# 3 credit class - got a B
# 2 credit Class - got a A
#######################################################################.
# ANSWER
gpa( credits = c(4,3,3,2), grades = c(2, 4, 3, 4))Error in gpa(credits = c(4, 3, 3, 2), grades = c(2, 4, 3, 4)): could not find function "gpa"##################################################.
# QUESTION
#
# Rewrite the definition of the gpa function, call it gpa2.
# This time, use the weightedAvg function inside
# your definition of the gpa function.
###################################################.
# ANSWER
gpa2 <- function (credits, grades) { # **ANSWER**
# **ANSWER**
totalCredits = sum(credits) # **ANSWER**
weights = credits / totalCredits # **ANSWER**
# **ANSWER**
# Instead of writing # **ANSWER**
# # **ANSWER**
# sum(weights * grades) # **ANSWER**
# # **ANSWER**
# as we did in the gpa function above, we can make use of the # **ANSWER**
# weightedAvg function that we defined earlier. # **ANSWER**
# # **ANSWER**
# Remember that we defined the weightedAvg function above as: # **ANSWER**
# # **ANSWER**
# weightedAvg = function(weights, tests) { # **ANSWER**
# sum ( weights * tests ) # **ANSWER**
# } # **ANSWER**
# # **ANSWER**
# NOTE that the weightedAvg function definition uses the # **ANSWER**
# argument names weights and tests # **ANSWER**
# In this function we want the weightedAvg of weights and grades. # **ANSWER**
# The following function call uses the weightedAvg function. # **ANSWER**
# The function all passes the weights variable that is local # **ANSWER**
# to the gpa2 function to the weights variable that # **ANSWER**
# is local to the weightedAvg function. However, the grades # **ANSWER**
# variable in this function is being passed to the # **ANSWER**
# "tests" variable in the weightedAvg function. # **ANSWER**
# **ANSWER**
weightedAvg( weights , grades ) # **ANSWER**
} # **ANSWER**
# calculate Joe's GPA. Joe took 4 classes. The credits and grades were as follows
# 4 credit class - got a C
# 3 credit class - got an A
# 3 credit class - got a B
# 2 credit Class - got a A
gpa2( credits = c(4,3,3,2), grades = c(2, 4, 3, 4))[1] 3.083333# CHECK YOUR ANSWER
# If you wrote the code correctly you should get the following results.
# Show that the gpa and the gpa2 functions both return the same results
gpa( credits=c(4,3,0.5) , grades=c(3,4,3)) # 3.4Error in gpa(credits = c(4, 3, 0.5), grades = c(3, 4, 3)): could not find function "gpa"gpa2( credits=c(4,3,0.5) , grades=c(3,4,3)) # 3.4[1] 3.4# Show that the gpa and the gpa2 functions both return the same results
gpa ( c(4,3,3,2) , c(2,4,3,4) ) # 3.08333Error in gpa(c(4, 3, 3, 2), c(2, 4, 3, 4)): could not find function "gpa"gpa2 ( c(4,3,3,2) , c(2,4,3,4) ) # 3.08333[1] 3.083333