CODEWITHSUNDEEP
rpattern-matchinggrepl
user2543622
user2543622

Reputation: 6796

R search for multiple patterns grepl

I have below code. I want to find cells that have alphanumeric values and it should also ignore the cells that are na or NA.

How can I modify my code to that? the required R command should return below result for newcolumn

True, True, false, false, True, false, false

I tried commands 3 and 4, but they failed :(

> newcolumn=c(1,2,"na","NA","abc","","*")

> grepl("[[:alnum:]]", newcolumn)
[1]  TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE

> grepl("[[:alnum:]] | na", newcolumn)
[1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE

> grepl(c("[[:alnum:]]","na"), newcolumn)
[1]  TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE
Warning message:
In grepl(c("[[:alnum:]]", "na"), newcolumn) :
  argument 'pattern' has length > 1 and only the first element will be used

> grepl("[[:alnum:]]" | "na" | "NA", newcolumn)
Error in "[[:alnum:]]" | "na" : 
  operations are possible only for numeric, logical or complex types

> str(newcolumn)
 chr [1:7] "1" "2" "na" "NA" "abc" "" "*"

===========================update1===============================

newcolumn2<-newcolumn[grepl("(?=(?i)na(N)?(*SKIP)(*F))|[[:alnum:]]|(?=(?i)nan(*SKIP)(*F))|(?=(?i)null(*SKIP)(*F))", newcolumn, perl=TRUE)]

I updated my code as above as i want to recognize na, nan, null and their variants too. But "null part is not working. What changes should i make?

Upvotes: 0

Views: 2056

Answers (1)

akrun
akrun

Reputation: 887881

Try:

 grepl("(?=(?i)na(*SKIP)(*F))|[[:alnum:]]", newcolumn, perl=TRUE)
 #[1]  TRUE  TRUE FALSE FALSE  TRUE FALSE FALSE

(?i) means case-insensitive. So it should match na, NA, nA, or Na. (*SKIP)(*F) in the pattern makes the match to fail. Now the pattern in the right side of | symbol ie. [[:alnum:]] would be the one that would be matched.

Update

 newcolumn <- c(1,2,"na","NA","abc","","*", "NaN", "nan", "nAn")
 grepl("(?i)na(N)?(*SKIP)(*F)|[[:alnum:]]", newcolumn, perl=TRUE)
 # [1]  TRUE  TRUE FALSE FALSE  TRUE FALSE FALSE FALSE FALSE FALSE

Upvotes: 1

Related Questions