在 R 中使用 grepl() 匹配句子中的两个连续单词（或：如何在 grepl() 中使用通配符）？

Question

Suppose I would like to match two consecutive words in a sentence but explicitly not match other sentences that might still contain both of these words but not one right after the other.

mydata <- data.frame(text=c("I like pizza, and a read a novel.", "I like novels."))

So, if I do this...

grepl("lik.*? novel.*?", mydata$text, perl=T, ignore.case=T)

...I get "[1] TRUE TRUE", while what I need is "FALSE TRUE" given that "like" in the first sentence doesn't refer to "novel".

Now, this might be a bad example, given that I could simply search for "Like novel.*?" without a wildcard for the first word, but suppose further that I need to use this wildcard for the first of the two words, too.

And connected to that: How would one match a word in a sentence with a wildcard in the middle of said word?

Example:

mydata<-data.frame(text=c("xxx abc xxx", "xxx azc xxx", "xxx a bc xxx"))

I would like to match words that start with "a" and end with "c" no matter what comes in between but the condition is that this must be one word. Currently, I get a "TRUE" even for the third line while what I would need is a match for the first two but not for the third:

grepl("a.*?c", mydata$text, perl=T, ignore.case=T)

Answer 1

If the words are consecutive

grepl("like\\b \\bnovel", mydata$text, perl=TRUE, ignore.case=TRUE)
#[1] FALSE  TRUE

and for second case, we can usse the word boundary ( \\\\b ) at the beginning and end of

grepl("\\ba\\w+c\\b", mydata$text, perl = TRUE, ignore.case = TRUE)
#[1]  TRUE  TRUE FALSE

Here the pattern to match is a word boundary ( \\\\b ) followed by character 'a', one or more characters ( \\\\w+ ) and 'c' followed by word boundary ( \\\\b )