Practice Set: Functions

length(), nrow(), ncol()

• length() is the same as getting the last element’s index position
• nrow() and ncol() tell you how many rows and columns there are in a data.frame, respectively

length(demographics$SubjectID)

## [1] 5

nrow(demographics)

## [1] 5

ncol(demographics)

## [1] 4

Note that these don’t work properly for any object

# length on a 2-d object just gives the 2nd dimension (columns)
# lots of people get this confused and think it should return the
# number of rows!
length(demographics)

## [1] 4

# nrow and ncol of a 1-d vector don't make sense...
# there is no such thing as columns or rows in a vector
nrow(demographics$SubjectID)

## NULL

ncol(demographics$SubjectID)

## NULL

factor() and as.factor()

Both of these convert a character vector into a factor. Factors are a special case of character where there is inherent meaning/grouping. Categorical variables.

(Note: For the life of me, I can’t figure out the difference between factor() and as.factor() 🤷‍♀ so you can think of them as interchangable.)

# Our 'Location' variable consists of states. We might want
# to think of this as a categorical variable, where each state
# is a meaingfully different category from the next state.

# Don't forget to re-assign the variable so that it stores your new
# factor!
demographics$Location <- factor(demographics$Location)
demographics$Location

## [1] Missouri Iowa     Missouri Idaho    Maine   
## Levels: Idaho Iowa Maine Missouri

# To double check, use the `class()` function
class(demographics$Location)

## [1] "factor"

The LEVELS of a factor are all the different categories. Here, we have 4 states. R will always put these in alphabetical order unless you tell it otherwise (hint: you can change this by modifying one of the arguments).

table() for quick counts

Especially good for factors & logicals!

# how many states are in each category?
table(demographics$Location)

## 
##    Idaho     Iowa    Maine Missouri 
##        1        1        1        2

# states x smoker table
table(demographics$Location, demographics$SmokerTF)

##           
##            FALSE TRUE
##   Idaho        0    1
##   Iowa         1    0
##   Maine        1    0
##   Missouri     1    1

cbind() and rbind() for adding columns and rows, respectively

These are for if you have a vector that you want to add to a data.frame. You should make sure there’s the same number of items – which changes depending on if you’re adding a row or a column!

Let’s add a column consisting of survey scores, on a scale of 1 to 10. Then we’ll add a row.

# first, let's add a column for surveys. there are currently
# 5 rows/observations, so we'll want 5 elements in this vector
survey <- c(6, 4, 10, 2, 8)

# now use cbind() to add it to your data.frame
# the 1st argument is the data.frame
# the 2nd argument is the thing you want to add to it
# (you're combining BOTH things)

# note that we re-assign the entire data.frame
demographics <- cbind(demographics, survey)

# print it out to view
demographics

##   SubjectID Location SmokerTF SubjectAge survey
## 1       ID1 Missouri     TRUE         20      6
## 2       ID2     Iowa    FALSE         18      4
## 3       ID3 Missouri    FALSE         32     10
## 4       ID4    Idaho     TRUE         25      2
## 5       ID5    Maine    FALSE         25      8

############

# now let's add a row. there are 5 columns, so we need to add an
# element that will match up with each column
newParticipant <- c("ID6", "Iowa", FALSE, 28, 5)

# now use rbind() to bind by rows
# the 1st argument is the data.frame
# the 2nd argument is the row you want to add to it. This makes it the LAST row
# (if you switched the order, it would become the FIRST row)
demographics <- rbind(demographics, newParticipant)

# print it out to view
demographics

##   SubjectID Location SmokerTF SubjectAge survey
## 1       ID1 Missouri     TRUE         20      6
## 2       ID2     Iowa    FALSE         18      4
## 3       ID3 Missouri    FALSE         32     10
## 4       ID4    Idaho     TRUE         25      2
## 5       ID5    Maine    FALSE         25      8
## 6       ID6     Iowa    FALSE         28      5

There is 1 slight issue with rbind(). Vectors will always take on the least specific class So newParticipant is considered a character vector because all of it’s elements could be thought of as characters, but not everything could be considered numeric or logical. This means that when we added it to the demographics data.frame, it made any column that wasn’t already a character into characters!

class(demographics$SubjectID)

## [1] "character"

class(demographics$Location) # factors are special characters

## [1] "factor"

class(demographics$SmokerTF)

## [1] "character"

class(demographics$SubjectAge)

## [1] "character"

class(demographics$survey)

## [1] "character"

However, the last 3 columns should NOT be characters. I change them back below. See if you can figure out what each function does.

demographics$SmokerTF <- as.logical(demographics$SmokerTF)
class(demographics$SmokerTF)

## [1] "logical"

demographics$SubjectAge <- as.numeric(demographics$SubjectAge)
class(demographics$SubjectAge)

## [1] "numeric"

demographics$survey <- as.numeric(demographics$survey)
class(demographics$survey)

## [1] "numeric"

(Ok for real this time…The End)