In stephlocke/RMSFTDP: R in the Microsoft Data Platform

Home | Prev | Next

Working with vectors {.tabset}

Intro

Vectors are the building blocks of R. They are 1D arrays that contain data of all the same type - so coercion of data types happens.

• R is a 1-based array system so a vector starts at position 1
• Items in a vector can have names but can also be referred to by position

Create

You can create vectors using the c() function or via a sequence (:). Values in a vector can be given names - this can be useful when subsetting values.

handCrafted<-c(1,2,3,4)
seqCrafted<-1:4
named<-c(a=1,b=2,c=3,d=4)
named

Filter

Alternatively called subsetting, we can filter a vector by using positive notations of position, negative notations of position, the name of a value, or providing a boolean value for each item in the vector.

handCrafted[1]
seqCrafted[-1]
named["b"]

handCrafted[c(rep(TRUE,3),FALSE)]

Update

You can update one or more values in a vector by assigning the new values into the desired subset.

handCrafted
handCrafted[2]<-99
handCrafted

named
named["a"]<-99
named

# Delete by subsetting without value
seqCrafted
seqCrafted<-seqCrafted[-4]
seqCrafted

# Append by creating a vector combining the original and the additional values
ordered
ordered<-c(ordered,5)
ordered

Manipulate

You can manipulate a vector using functions. You can overwrite any object created - this will change mode, class, etc as R is loosely typed and doesn't require such things to be specified and fixed up front.

mode(seqCrafted)
seqCrafted<-as.character(seqCrafted)
mode(seqCrafted)

Order

Ordering of records works by providing the position numbers of values in a vector and then using those to produce a vector with the original components in new locations

preOrder<-sample(letters, 6)
preOrder

# Get order the values should appear in to be alphabetised
order(preOrder)

# Use it to sort a vector
ordered<-preOrder[order(preOrder)]
ordered

# Alternatively, use the sort() function for brevity
sorted<-sort(preOrder)
sorted

Metadata

You can extract various pieces of information about a vector

names(handCrafted)
names(named)

dim(named)
dimnames(named)

length(named)
class(named)
mode(named)

attributes(named)
str(named)

Exercises

1. Create a vector containing upper case and lower case variants of the alphabet
2. Create a new vector with a random sample of your new letter vector
3. Filter out any lowercase letters

Answers

#1
lets<-c(LETTERS,letters)
#2
lets<-sample(lets,50)
#3
lets<-lets[tolower(lets)!=lets]
lets

Working with lists {.tabset}

Intro

Lists hold multiple objects together and form the basis of complex data objects like data.frames and model results.

Create

You can create lists using the list() function or via a sequence (:). Objects in a list can be given names.

basicList<-list(c(1,2,3,4),LETTERS[5:8], rnorm(5))
namedList<-list(p1=c(1,2,3,4),p2=LETTERS[5:8])

Filter

Alternatively called subsetting, we can filter a list by using positive notations of position, negative notations of position, or the name of a element, or providing a boolean value for each item in the list. We can also use list[[ "elementname" ]] or `list$elementname for specifically detailing a single element.

basicList[1]
basicList[-1]
namedList["p2"]
basicList[c(TRUE,FALSE)]

basicList[[1]]
basicList[[-1]]
namedList[["p2"]]
namedList$p2
basicList[[c(TRUE,TRUE)]]

Update

You can update one or more objects in a list by assigning the new values into the desired object using the same subsetting capabilities as noted in the Filter section.

basicList[[1]]<-8:12
basicList[1]

# Elements in a list can be removed by making them NULL
basicList[2]
basicList[2]<-NULL
basicList[2]

# Append by creating a list combining the original and the additional values
basicList[[3]]<-LETTERS[5:8]

Manipulate

You can manipulate a list using functions, and also manipulate the objects stored in the list.

lapply(basicList,mode)

Order

Ordering of objects in a list is rarely required but you can do it with the order() function.

unorderedList<-list(p2=c(1,2,3,4),p1=LETTERS[5:8])
unorderedList[order(names(unorderedList))]

Metadata

You can extract various pieces of information about a list

names(basicList)
names(namedList)

dim(namedList)
dimnames(namedList)

length(namedList)
class(namedList)
mode(namedList)
attributes(namedList)
str(namedList)

Exercise

Create a linear regression model (lm()) for the iris dataset and extract the fitted.values element

Answers

irisLM<-lm(Sepal.Width~Sepal.Length, iris)
head(irisLM$fitted.values)

Working with tables {.tabset}

Intro

Tables or data.frame's as the base structure is called in R can hold multiple columns of different data types. Normally, data.table or dplyr would be taught to super-charge data.frames (see Steph's extended session "Cut the R learning curve" for more on these) but to reduce dependencies and make it easier to transfer code between different systems, we'll use just base R.

• data.frames have a coordinate system like excel so df[ 1 , 2 ] selects the intersection of row 1 and column 2
• Rows and columns can be referenced by name
• A data.frame is actuallly a list that is presented like a table

Create

You can create data.frames using the data.frame() function.

• data.frame() will error if you try storing something odd in it. Use as.data.frame() to coerce

df<-data.frame(a=1:4, b=LETTERS[5:8], c=rnorm(4),row.names = letters[9:12])
df

Filter

Alternatively called subsetting, we can filter a data.frame by using positive notations of position, negative notations of position, or the name of a element, or providing a boolean value for each item in the list.

To filter using a condition requires the creation of a boolean vector based on a specific column. columns can be treated as vectors by using df$colname

df[1, ]
df[ ,1]
df[1,1]
df[-(3:4),]
df[,"a"]
df[ , c(TRUE, TRUE, FALSE)]
df[df$a<4,]

Update

You can update values in a data.frame by referencing it's position using the same subsetting capabilities as noted in the Filter section.

df[1,1]<-2
df

# Columns in a data.frame can be removed by making them NULL
df[,2]
df[,2]<-NULL
df[,2]

# Rows in a data.frame can be removed by subsetting without them
df[2,]
df<-df[-2,]
df[2,]

# Appends can happen in a variety of ways
superDF<-data.frame(df,d=5:7)
superDF
df$newcol<-5:7
df
df[4,]<-c(1,1,1)
df

Order

Ordering of data.frames works by providing the position numbers of row numbers in a vector and then using these to return data.frame rows in a specific order

# Get order the values should appear in
order(df$c)

# Use it to sort a table
ordered<-df[order(df$c),]
ordered

Metadata

You can extract various pieces of information about a list

names(df)

dim(df)
dimnames(df)

length(df)
class(df)
mode(df)
attributes(df)
str(df)

Exercises

1. Make a copy of the iris dataset to play with
2. Add a column with the estimated area of the sepals
3. Filter out any record with a petal width lower than average
4. Sort by species name in descending order

Answers

#1 
myIris<-iris
#2
myIris$Sepal.Area<-myIris$Sepal.Width * myIris$Sepal.Length
#3
avg<-mean(myIris$Petal.Width)
myIris<-myIris[myIris$Petal.Width>=avg,]
#4
myIris<-myIris[order(myIris$Species,decreasing = TRUE),]
head(myIris)

stephlocke/RMSFTDP documentation built on May 30, 2019, 3:36 p.m.