SessionTwo_MakingFunctionCalls

Session Two - Making Function Calls
By Hellen Gakuruh
May 21, 2016
Table of Contents
Session Goal................................................................................................................................................................2
What we shall cover ...............................................................................................................................................2
Prerequisite................................................................................................................................................................2
R Functions.................................................................................................................................................................2
Making Function Calls ...........................................................................................................................................3
Function's composition....................................................................................................................................3
A functions name ................................................................................................................................................3
Arguments..............................................................................................................................................................4
Named and Unnamed Argument............................................................................................................4
Position matching for named and unnamed arguments.............................................................5
Default values for arguments...................................................................................................................6
Gauge yourself...........................................................................................................................................................8

2
Session Goal
The main goal of this session is to equip you with the necessary skills to make function calls
in R.
What we shall cover
By the end of this session you should:
• Be conversant with the term function and function call
• Know what constitutes a function and a function call
• Be conversant with arguments
• Distinguish between named and unnamed arguments
• Know how R positions arguments in a call
• Be able to know and change default values
• Have an understanding of what top and low level functions are
• Know what is meant by anonymous functions
Prerequisite
To appreciate this session,
• you must have R and RStudio installed and running
• know how to use the console
• be conversant with RStudio panes like console, help and packages
R Functions
A function is a command you give to a computer program; basically an action you what
performed like getting the sum/total, mean or determine existence of a file in your
computer. There is virtually nothing that can happen in R without the use of a function, for
example, if you want to get data into R, or clean it (data manipulation and transformation),
or analyze it, or even export its output, you must use a function.
Presence of parenthesis “( )” like sum( ), mean( ) or file.exist( ) indicated a function and
when it is used, it is referred to as making a function call.
Before we discuss function calls, let's note the two types of functions in R; these are in-built
functions like sum, mean as well as file.exist and those you develop (also known as user
defined functions). In this session we will discuss in-built functions that come with base R
in form of (default) packages and those available from add-on packages. User defined
functions will be covered in level 2.

3
Making Function Calls
Function's composition
First thing you need to know is the composition of function. A function in R is composed of
a name, arguments/parameters, and Body.
Function composition
A function begins with its name followed by parenthesis and curly braces (body). To make
a function call you need a function's name and it's parenthesis ( ), for example mean( ) (the
body or what is in curly braces is not necessary to make a call but it will give you insight as
to how the function works).
A functions name
Most functions in R are named after the actions they should perform, for example the two
average functions mean and median. Others include sum, exit, ncol (number of columns),
nrow (number of rows) e.t.c.
However, there are numerous other functions that might not follow this kind of logic, for
example, you might expect a function called "mode" to get the most frequently occurring
observation (the third measure of central tendency), however, this is not the case as mode
in R has a totally different meaning. It gives the internal type of an object (we shall discuss
this in the fourth session)
Key point to note here is that, knowing a function's names comes with practice, reading R
manuals (help.start()) and online tutorials/discussions/Q and A.
One final tip on function names, there are unnamed functions usually included in other
functions, these are referred to as anonymous functions.
For example, the function within the sapply function (with one argument x) is an
anonymous function.
a <- list(1:10, 11:20, 21:30, 31:40, 41:50)
b <- list(round(rnorm(10), 1), round(rnorm(10), 1), round(rnorm(10), 1),
round(rnorm(10), 1), round(rnorm(10), 1))

4
# Anonymous function
sapply(1:5, function(x) a[[x]] + b[[x]])
## [,1] [,2] [,3] [,4] [,5]
## [1,] 0.9 9.7 21.9 31.5 41.1
## [2,] 2.3 11.7 21.5 29.9 42.9
## [3,] 2.1 14.6 22.5 34.0 42.2
## [4,] 2.9 14.2 23.0 32.7 43.6
## [5,] 6.2 14.3 25.2 35.7 45.3
## [6,] 6.3 14.4 26.0 36.4 45.4
## [7,] 6.6 17.8 26.4 36.7 49.1
## [8,] 6.5 18.5 28.3 37.5 46.1
## [9,] 9.8 19.3 29.7 38.7 48.5
## [10,] 11.4 19.0 28.6 41.0 50.2
Arguments
Within the parenthesis of a function are arguments or parameters. These arguments
specify how a function will work or how a command will be carried out. It's important to
know a function's arguments as they have implications on the output you receive.
Sometimes this can result is an error meaning nothing is produced or a warning.
To know a function's argument, you need to read its documentation which can be accessed
with the function help() or preceding the function's name with a questions mark. For
example, get documentation for the mean function by typing either help(mean) or ?mean
on the console. The documentation should appear on the help tab located on the right
lower pane.
From this documentation (at the top), you will notice the name of the function requested,
and in curly braces the package to which the documentation comes from i.e. base package.
The package name is given to identify location of a function. This stems from the fact that
other non-base R packages can have similar names as those in base R.
Right below is the title of the documentation, in this case the arithmetic mean; R's default
mean. A short description of the function is also given followed by a section on the
function's usage. This "Usage" section shows how the function is called in terms of its name
and arguments.
But before we discuss this section, there are some critical issues we need to discuss first,
these include naming of arguments, their order in a call and default values.
Named and Unnamed Argument
Going back to the mean documentation and specifically the usage section, we see two ways
of calling the mean function, that is R's mean(x, ...) and S3's mean(x, trim = 0, na.rm =
FALSE, ...). Both of these work the same, the minor difference can be understood by
reading up on Object Oriented Programming in R.
Looking at the second function, the S3 method, there are four arguments, these are x, trim,
na.rm, and .... These are the names of the arguments and to make a call you need to give
values to these arguments.

5
You can make a function call with only values specified or with named arguments. To
understand what this means, let's first get to know what the arguments for the mean
function are and their expected values: The section called Arguments will guide us on this.
From this section we can see that x is an object (object can be described as any entity in R)
and the expectation is that it should either be a numeric/logical vector (variable), date,
date-time or time interval objects: Anything other than this will raise a warning. Trim is a
fraction of observations to remove from computation (suitable for dealing with skewed
data or outliers ), the expectation is that it should be a value between 0 and 0.5. In R, NA
means missing data and therefore na.rm is an argument indicating how missing data
should be handled, the expected values are either true or false (remove or don't remove).
For now let's leave the ... argument as it requires a bit of knowledge in methods.
Now that we know what x, trim and na.rm mean, lets look at how to make function calls
with both named and unnamed argument.
# To call the mean function, you can name the arguments.
mean(x = c(2, 5, 9, 3), trim = 0.5, na.rm = TRUE)
## [1] 4
# Or call with unnamed arguments
mean(c(2, 5, 9, 3), 0.5, TRUE)
## [1] 4
Position matching for named and unnamed arguments
If you call a function with named arguments, then the position of the arguments do not
matter.
# Matching position of the arguments.
mean(x = c(2, 5, 9, 3), trim = 0.5, na.rm = TRUE)
## [1] 4
# Mixing the order of arguments
mean(trim = 0.5, x = c(2, 5, 9, 3), na.rm = TRUE)
## [1] 4
But if you have unnamed arguments then you need to give the values of arguments
according to their order as R will match each value with argument in that position.
# Mixing order for unnamed argument
mean(TRUE, 0.5, c(2, 5, 9, 3))
## [1] TRUE
## Warning message:
## In if (na.rm) x <- x[!is.na(x)] :
## the condition has length > 1 and only the first element will be used
R has given us a result of TRUE, not exactly what we expected, and in addition it has given
us a warning indicating a problem with the call. What happened is that R marched TRUE to
the X argument, 0.5 to the trim argument and c(2, 5, 9, 3) to the na.rm argument. Since the

6
expected values for X include logical values, putting true where x should be is okay by R,
however, when you put more than one value where one value (true/false) is expected, then
you will get a warning message.
If you have named and unnamed arguments the rule of the thumb is to name subsequent
arguments, otherwise R will match the values with the position of the other arguments.
# Naming the second and subsequent argument
mean(c(2, 5, 9, 3), trim = 0.5, na.rm = TRUE)
## [1] 4
# Naming only the second argument
mean(c(2, 5, 9, 3), trim = 0.5, TRUE)
## [1] 4
The second call did not raise a warning as the last value was matched to the correct
argument but this might not be the case in other situations.
Now let's look at default values.
Default values for arguments
Arguments can have default values which are used if a call is made without the argument.
For example, the mean function has two arguments with default values, these are trim and
na.rm. This means a call can be made with only the value of x, but if x is missing, then you
will get an error.
# Calling the mean with only x
mean(x = c(2, 5, 9, 3))
## [1] 4.75
# You call with default values but that's just unnecessary
mean(x = c(2, 5, 9, 3), trim = 0, na.rm = FALSE)
## [1] 4.75
# Making a function call without the x value
mean(trim = 0.5, na.rm = TRUE)
##Error in mean.default(trim = 0.5, na.rm = TRUE) :
## argument "x" is missing, with no default
If a default value is not what you want, you can add the argument specifying the value you
want. For example, since the above data set has an extreme value (9), we can use the trim
function to make the average reasonable.
# Trimming 0.5 (50%) from the dataset or 0.25 (25%) from beginning and the
end of the dataset instead of the default 0
mean(c(2, 5, 9, 3), trim = .5)
## [1] 4
# Compare with the median

7
median(c(2, 5, 9, 3))
## [1] 4
If we had a data set with missing values, then we need to tell R to make the computation
without the missing values otherwise it will result in NA
# Taking the mean of a dataset with NA when na.rm is false (default)
mean(c(2, 5, 9, 3, NA), trim = .5)
## [1] NA
# Taking the mean of a dataset with NA when na.rm true
mean(c(2, 5, 9, 3, NA), trim = .5, na.rm = TRUE)
## [1] 4
The mean has a short and straightforward list of arguments, but others might have a
number of arguments. For example, pull up the help documentation for importing data;
that is, help(read.table). This function has more than 20 listed arguments. But if you take
a keen look at most of the arguments, you will notice that most of the arguments have
default value. It is of great interest to you to read through the descriptions of all the
arguments and confirm that the defaults are okay, this is usually a source of many
challenges in function calls and expected outputs.
In addition to checking the help documentations, you can use the function args() to see the
list of arguments for a specified function.
# Mean arguments
args(mean)
## function (x, ...)
## NULL
# Read.table arguments
args(read.table)
## function (file, header = FALSE, sep = "", quote = ""'", dec = ".",
## numerals = c("allow.loss", "warn.loss", "no.loss"), row.names,
## col.names, as.is = !stringsAsFactors, na.strings = "NA",
## colClasses = NA, nrows = -1, skip = 0, check.names = TRUE,
## fill = !blank.lines.skip, strip.white = FALSE, blank.lines.skip =
TRUE,
## comment.char = "#", allowEscapes = FALSE, flush = FALSE,
## stringsAsFactors = default.stringsAsFactors(), fileEncoding = "",
## encoding = "unknown", text, skipNul = FALSE)
## NULL
It is also useful to note that there are functions called top level and those that are called
low level. Top level functions are functions that we mostly use to perform an action, while
low level functions are called by the top level functions. For example, when computing the
mean, you would call on the mean() function which in-turn calls the sum and length
function to get you the mean. So in this case the mean is a top level function while sum()
and length() are the low level functions or helper functions.

8
Later in level two as we learn user defined functions, we shall discuss functions in terms
of argument, body and environment.
Gauge yourself
1. What are functions?
2. How can you tell an object is a function?
3. What is meant by a function call?
4. What constitutes a function call?
5. What are arguments?
6. Distinguish between named and unnamed arguments
7. What do you expect would happen if I made a function call with all the necessary
argument but only the second argument is named and distort the position of the other
arguments?
You are ready for the third session if you know these
1. Functions are commands issued for an action to be performed
2. In R an object name followed by parenthesis indicates a function
3. A function call occurs when a function is used or run
4. A function call constitutes a name of a function and it's arguments
5. Arguments give details or input for function execution
6. A function call with arguments and their values are referred to as named arguments
while a call with values only is known as unnamed arguments.
7. Since all the arguments are provided, then you will not receive an error, however, due
to distortion of position, if the value given does not meet the expected value, a warning
will be issued and output generated will be questionable.
Note
This write-up has used non programming definition in as much as possible with the intent
of being useful to non-programmers and imparting concepts and skills necessary for
analysis.

SessionTwo_MakingFunctionCalls

More Related Content

What's hot (14)

Viewers also liked (13)

Similar to SessionTwo_MakingFunctionCalls (20)

More from Hellen Gakuruh (20)

SessionTwo_MakingFunctionCalls