R : regular expression to match pattern in only the first line [closed] - regex

Question

Closed. This question needs details or clarity. It is not currently accepting answers.
Want to improve this question? Add details and clarify the problem by editing this post.
Closed 8 years ago.
Improve this question
In my R code, I have the following content of x as a result of lda prediction output.
[1] lamb
Levels: lamb cow chicken
I would like to capture the word "lamb" in the first line and not the second line.
I had the following reg expression which did not work.
if (regmatches(x,regexec(".*?([a-z]+)",x))[[1]][2]=="lamb"){
cat("It is a lamb")
}
Instead, I also got the following error :-
Error in regexec(".*?([a-z]+)", x) : invalid 'text' argument
Anyone with help ?
Thanks in advance.
mf

Answer 1

Direct Answer:
It is a variable type error. See ?predict.lda to learn why: The return object of a predict() when used with an object of class lda is a list. You just want the first element of the list, which is a factor for an object of type integer. Factors in R store some characters for every element in their level component, which can be accessed by levels() (Read ?factor as well.). But what you want is to access the explicit value your factor shows, which can be acheived by as.character(). By the way: The second line does not get checked by the regex. It is just standard console output of a factor, see ?print.factor.
Here's an example, based on thepredict.lda() help page:
tr <- sample(1:50, 25)
train <- rbind(iris3[tr,,1], iris3[tr,,2], iris3[tr,,3])
test <- rbind(iris3[-tr,,1], iris3[-tr,,2], iris3[-tr,,3])
cl <- factor(c(rep("s",25), rep("c",25), rep("v",25)))
z <- lda(train, cl)
x_lda <- predict(z, test)
# x_lda is a list
typeof(x_lda)
# The first element of the list, called "class", is a factor of type integer.
typeof(x_lda$class)
# Now we create a character vector from the factor:
as.character(x_lda$class)
With an explicit character object, your code works for me:
x <- "lamb"
regmatches(x,regexec(".*?([a-z]+)",x))[[1]][2]=="lamb"
[1] TRUE
So you need to coerce your object to character, and then use it as the "text" argument for the regexec function.
Actual Answer:
There are better ways to do this.
You nest and chain a lot of functions in one line. This is barely readable and makes debugging hard.
If you know that the output will always consist of certain elements (especially, since you know the input of your lda prediction and therefore know the different factor levels beforehand), you can simply check them by == and maybe any() (continuing with the example from before):
levels(cl)
[1] "c" "s" "v"
any(as.character(x_lda$class)=="c")
[1] TRUE
See the help file for ?any, if you don't know what it does.
Finally, if you just want to print "It is a lamb" in the end, and your output will always just have one element, you can simply use paste():
paste("It is a", as.character(x))
[1] "It is a lamb"