Check if a string is a date with a regex but the result is not the expected one

[hich]

Assume the following list:

string_test <- c("na-05-2021", "nk-31-2006", "NK-nk-1965", "NK-nk-89", "00-12-2005", "12-12-nk")

I would like to check if the strings respect the following date format: dd/mm/YYYY
If the day or the month is unknown, we admit to replace it by nk (upper or lower case).
So, the accepted strings in our example are: "na-05-2021", and "NK-nk-1965".
The other are not accepted.

I did the following regex but the returned results are not what i expected:

#install.packages("stringr")
library(stringr)
result <- str_detect(string_test , "/^(nk|NK|(0[1-9]|[12]/d|3[01]))+(-)+(nk|NK|0[1-9]|1[0-2])(-)+(1|2)+(0|9)+[0-9]+[0-9]/", negate = TRUE)

I think it's a problem with the regex but i tested it in https://regexr.com/ and it works as expected.

[andresrcs]

The problem is that you are using Javascript regex flavor and stringr uses ICU flavor, for example, in ICU you use ^ instead of \^

[hich]

Sorry this is my correct code:

str_detect(test_string, "^(nk|NK|((0[1-9]|[12]\\d|3[01])))+(-)+(nk|NK|(0[1-9]|1[0-2]))(-)+((1|2)+(0|9)+[0-9]+[0-9])")

You are right but where can I find a website to test my regex in ICU flavor because I didn't find anything and also, I don't understand why it doesn't give me an error but an incorrect result.

[andresrcs]

@hich

where can I find a website to test my regex in ICU flavor

You can test it from within R with regexplain

GitHub - gadenbuie/regexplain: 🔍 An RStudio addin slash regex utility belt

@hich

I don't understand why it doesn't give me an error but an incorrect result.

Because you are scaping the metacharacter so it is not an invalid regex expression, it is just looking for a literal "^"

[hich]

Thank you, i found on regexplain the following regex to test the correct format date:

(0?[1-9]|[12][0-9]|3[01])([ /-])(0?[1-9]|1[012])\2([0-9][0-9][0-9][0-9])(([ -])([0-1]?[0-9]|2[0-3]):([0-5]?[0-9]):[0-5]?[0-9])?

But it checks only if the date is in the format dd-mm-yyyy, so I tried to add the "nk" condition but nothing works.

Have you an idea of how add this condition please ? or anyone else

[hich]

I finally found the correct regex to do it, I put it here if anyone has the same problem:

^(nk|NK|(0[1-9]|[12][0-9]|3[01]))(-)(nk|NK|(0[1-9]|1[0-2]))(-)((1|2)(0|9)[0-9][0-9])

It test if the string match with a correct date format dd-mm-yyyy and if day or month is unknown, "nk" (upper or lower case) is admitted. (e.g.: nk-01-2021, nk-nk-2021, etc.)