Working with Text

Working with Text

# Count number of characters in a bit of text: 
sentence <- "Not a very long sentence." 
nchar(sentence)

## [1] 25

# Extract the first 3 characters: 
substr(sentence, 1, 3)

## [1] "Not"

# Substring all elements of a vector 
substr(c("good","good riddance","good on ya"),1,4)

## [1] "good" "good" "good"

# Number of characters of all elements of a vector 
nchar(c("hey","hi","how","ya","doin"))

## [1] 3 2 3 2 4

# Add a suffix to each text element of a vector: 
txt <- c("apple","pear","banana") 
paste(txt, "-fruit")

## [1] "apple -fruit"  "pear -fruit"   "banana -fruit"

# Glue them all together into a single string using the collapse argument 
paste(txt, collapse="-")

## [1] "apple-pear-banana"

# Combine numbers and text: 
paste("Question", 1:3)

## [1] "Question 1" "Question 2" "Question 3"

# This can be of use to make new variables in a dataframe, 
# as in this example where we combine two factors to create a new one: 
dataset <- read.csv("att.csv")
dataset$campaign <- with(dataset, paste(reachout, card, sep="-")) 
head(dataset$campaign)

## [1] "NO-NO"  "NO-NO"  "NO-YES" "NO-NO"  "NO-NO"  "NO-NO"

Column names

# Change the names of a dataframe: h
names(dataset) # first print the old names

##  [1] "pick"       "income"     "moves"      "age"        "education" 
##  [6] "employment" "usage"      "nonpub"     "reachout"   "card"      
## [11] "campaign"

names(dataset) <- c("pick1","income1","moves1","age1","education1", "employment1","usage1","nonpub1", "reachout1","card1") # then change the names

# Change only the first name (you can index names() just like you can a vector!) 
names(dataset)[1] <- "provider" 

#???nd out which columns have particular names
match(c("reachout1","card1"), names(dataset))

## [1]  9 10

Text in dataframes and grep

When reading a dataset, R converts automatically any numbers into a numeric, otherwise a factor. Sometimes we want to keep a variable to be as a text. The arguement “stringAsFactors=FALSE” avoided the automatic conversion of character variables to factors:

# Read data, tell R to treat the last variable ('card') as character, not factor 
dataset <- read.csv("att.csv") 
dataset$card <- as.character(dataset$card)

# Read data, tell R to treat all non-numeric variables as character
dataset <- read.csv("att.csv", stringsAsFactors=FALSE)

# Make sure that the Pick is really a character vector: 
is.character(dataset$pick)

## [1] TRUE

#Check the types of all variables in the dataset
print(str(dataset))

## 'data.frame':    1000 obs. of  10 variables:
##  $ pick      : chr  "OCC" "ATT" "OCC" "OCC" ...
##  $ income    : chr  "<7.5" "45-75" "" "" ...
##  $ moves     : chr  "0" "2" "0" "2" ...
##  $ age       : chr  "35-44" "25-34" "" "65+" ...
##  $ education : chr  "HS" "HS" "" "<HS" ...
##  $ employment: chr  "F" "F" "" "R" ...
##  $ usage     : int  9 2 6 7 0 0 3 1 0 2 ...
##  $ nonpub    : chr  "YES" "YES" "NO" "NO" ...
##  $ reachout  : chr  "NO" "NO" "NO" "NO" ...
##  $ card      : chr  "NO" "NO" "YES" "NO" ...
## NULL

#Create new variable to descibe for the next function
dataset$campaign <- with(dataset, paste(reachout, card, sep="-"))
dataset$campaign <- paste(dataset$campaign, 1:1000)

# Extract campaign (for convenience). 
campaign <- dataset$campaign

# Find the observations that have "YES-YES" in them. 
# grep() returns the index of values that contain 'YES-YES' 
grep("YES-YES",campaign)

##  [1]  31 235 362 410 424 441 471 484 489 537 552 575 599 604 605 625 626
## [18] 678 721 749 770 773 793 817 845 883 887 891 931 935

#grepl("YES-YES",campaign) # to returns TRUE or FALSE 

# That result just gives you the indices of the vector that have "YES-YES" in them. 
# these are the corresponding names: 
campaign[grep("YES-YES",campaign)]

##  [1] "YES-YES 31"  "YES-YES 235" "YES-YES 362" "YES-YES 410" "YES-YES 424"
##  [6] "YES-YES 441" "YES-YES 471" "YES-YES 484" "YES-YES 489" "YES-YES 537"
## [11] "YES-YES 552" "YES-YES 575" "YES-YES 599" "YES-YES 604" "YES-YES 605"
## [16] "YES-YES 625" "YES-YES 626" "YES-YES 678" "YES-YES 721" "YES-YES 749"
## [21] "YES-YES 770" "YES-YES 773" "YES-YES 793" "YES-YES 817" "YES-YES 845"
## [26] "YES-YES 883" "YES-YES 887" "YES-YES 891" "YES-YES 931" "YES-YES 935"

# Now find the cereals whose name starts with "YES-YES 3". 
# The ^ symbol is part of a 'regular expression', it indicates 'starts with': 
grep("^YES-YES 3",campaign)

## [1]  31 362

# Or end with "29"
# The $ symbol is part of a 'regular expression', and indicates 'ends with': 
grep("29$", campaign)

##  [1]  29 129 229 329 429 529 629 729 829 929

# grepl will return FALSE when "YES-YES" is not found, TRUE otherwise 
dataset$twoCampaign <- grepl("^YES-YES", campaign)

# Quick summary: 
summary(dataset$twoCampaign)

##    Mode   FALSE    TRUE 
## logical     970      30