有没有人可以使用 R 解释有关基本正则表达式的结果？

Question

1.

test1
# [1] "abcd" "abc1" "ab2b" "a3cd" "4bcd"    
test1[grep('[^abc1-3]', test1)]
# [1] "abcd" "a3cd" "4bcd"
test1[grep('[^a-d1-3]', test1)]
# [1] "4bcd"
test1[grep('[^4]', test1)]
# [1] "abcd" "abc1" "ab2b" "a3cd" "4bcd"

2.

test5 <- c('lo', 'lol', 'lolo', 'olo', 'lool')
test5[grep('loll*', test5)]
# [1] "lol"  "lolo"

3.

test5
# [1] "lo"   "lol"  "lolo" "olo"  "lool"
test5[grep('lolo+', test5)]
# [1] "lolo"
test5[grep('lol+', test5)]
# [1] "lol"  "lolo"

I'm studying basic contents about regular expression using R. But I can't understand why above three examples return those results.

For example, when using ^ in [], I learned that it returns the characters which don't involve the letters behind ^. But results don't seem like that.

I'm not good at English so I have difficulty in explaining all kinds of thing that I can't understand but I'll really appreciate if anyone can teach me why those results are returned by those R codes.

:(

Answer 1

You are correct that "[^abc1-3]" will match any character that is not in {a, b, c, 1, 2, 3}.

grep will return TRUE if there is any match, and FALSE if there is no match.

test1
# [1] "abcd" "abc1" "ab2b" "a3cd" "4bcd"    
test1[grep('[^abc1-3]', test1)]
# [1] "abcd" "a3cd" "4bcd"

The three results have d , which is not in {a, b, c, 1, 2, 3} , so d is matched (the last one has 4 , which is also matched). The two test items that are not in the results, "abc1" and "ab2b" only have characters in {a, b, c, 1, 2, 3} , so there is no match.

Regex101 is a good site for testing regular expressions and seeing how they work. Here is this example: https://regex101.com/r/CaxfCI/1

For your other examples,

• * means 0 or more . So 'loll*' matches lol followed by 0 or more l .
• + means 1 or more . So 'lol+' matches lo followed by 1 or more l .

Note, for grep , there is not any point in 'loll*' ... the result will be the same as for 'lol' . But in other regex operations, if you are replacing (substituting) or extracting the matches, the difference can matter.