I have the data frame "a" and it has a variable called "VAL". I want to count the elements where the value of VAL is 23 or 24.
I used two codes which worked Ok:
nrow(subset(a,VAL==23|VAL==24)
nrow(subset(a,VAL %in% c(23,24)))
But, I tried other code which gives an unexpected output and I don't know why.
nrow(subset(a,VAL ==c(23,24)))
Even if I change the order of 23 and 24, it gives a different unexpected output.
nrow(subset(a,VAL ==c(24,23)))
Why are those codes incorrect ? What are they actually doing?
Working through an example shows where it is going wrong:
a <- data.frame(VAL=c(1,1,1,23,24))
a
# VAL
#1 1
#2 1
#3 1
#4 23
#5 24
These work:
a$VAL %in% c(23,24)
#[1] FALSE FALSE FALSE TRUE TRUE
a$VAL==23 | a$VAL==24
#[1] FALSE FALSE FALSE TRUE TRUE
The following doesn't work due to vector recycling when comparing - take note of the warning message below E.g.:
a$VAL ==c(23,24)
#[1] FALSE FALSE FALSE FALSE FALSE
#Warning message:
#In a$VAL == c(23, 24) :
# longer object length is not a multiple of shorter object length
This last bit of code recycles what you are testing against and is basically comparing:
c( 1, 1, 1, 23, 24) #to
c(23, 24, 23, 24, 23)
...so you don't get any rows returned. Changing the order will give you
c( 1, 1, 1, 23, 24) #to
c(24, 23, 24, 23, 24)
...and you will get two rows returned (which gives the intended result by pure luck, but it is not appropriate to use).
