# notes on lapply()

## usage

lapply(X,FUN)

## arguments

X is a vector or list to loop over

FUN is the name a function to apply to each element of X

### lapply() basic example

fruits<-c("oranges", "bananas", "cherries")

iLike <- function(thingIlike) {
paste("I like", thingIlike)
}

lapply(fruits, iLike)

[[1]]
[1] "I like oranges"

[[2]]
[1] "I like bananas"

[[3]]
[1] "I like cherries"


### lapply() basic example explained

fruits<-c("oranges", "bananas", "cherries")

iLike <- function(thingIlike) {
paste("I like", thingIlike)
}

lapply(fruits, iLike)

• fruits is a simple vector with three elements
• iLike is a function with single argument called thingIlike
• the paste() function is used to combine the text “I like” whatever value we assign to the thingIlike argument
• we use lapply to loop over each element of fruits and apply the function to each in sequence

### lapply() in the homework

You were asked to use lapply() to loop over the iris dataframe and do a shapiro.test on each column.

You almost all did this perfectly.

iris_results <- lapply(iris[-5], FUN=shapiro.test)

iris_results[[1]]


Shapiro-Wilk normality test

data:  X[[1L]]
W = 0.9761, p-value = 0.01018


### lapply() in the homework - the offender

Next, you used lapply() on iris_results to extract the p-values. About half (or more) of you gave me this code.

lapply(iris_results, '[[', 2)

$Sepal.Length [1] 0.01018116$Sepal.Width
[1] 0.1011543

$Petal.Length [1] 7.412263e-10$Petal.Width
[1] 1.680465e-08


### lapply() in the homework - the offender

Some people also gave me this equivalent version

lapply(iris_results, '[[', "p.value")

$Sepal.Length [1] 0.01018116$Sepal.Width
[1] 0.1011543

$Petal.Length [1] 7.412263e-10$Petal.Width
[1] 1.680465e-08


### lapply() in the homework - a better way

Writing a function to pull out p values is MUCH easier to read!

get.p <- function(testresults) {
testresults$p.value } lapply(iris_results, get.p)  $Sepal.Length
[1] 0.01018116

$Sepal.Width [1] 0.1011543$Petal.Length
[1] 7.412263e-10

\$Petal.Width
[1] 1.680465e-08


### lapply() in the homework - my concern

many of you used the identical and very strange usage of '[[', which I didn't tell you about, and was very different from all the lapply() examples I have shown you.

## good news

somebody followed the JFGIS principle and found this way to get the pvalue!

It seems clear that many of you uncritically got this usage from said googler, and didn't think about the much clearer way to do it – one that was similar to other examples of lapply() I have given you–