Reputation: 2877
I wish to drop any rows in my data frame which do not meet the regular expression in a specific column i.e the cell must begin with two characters followed by four numbers, after that I do not care if it's U09 or U21. The first 6 characters are all that matters.
I am using the following code but I am getting 0 rows returned and I am not sure why:
with(prachData, prachData[grepl("^[A-Z][A-Z][0-9]{4}$", WCEL.name), ])
When I type head(prachData$WCEL.name)
I get the following details:
> head(prachData$WCEL.name)
[1] 0 0 CE0001U21B2 CE0001U21A3 CE0001U21C1 CE0001U21B1
13684 Levels: 0 1 11 12 13 2 21 22 23 3 31 32 33 CE0001U09A3 CE0001U09B3 CE0001U09C3 CE0001U21A1 CE0001U21A2 ... WX0114U09C3
And using class(prachData$WCEL.name)
I get:
[1] "factor"
Can anyone guide me to my mistake?
Upvotes: 1
Views: 795
Reputation: 887891
The problem seems to be the use $
in the pattern which means the end of string as it is a metacharacter, but based on the input showed, that is not the case because after the 4 digits there are other characters as well i.e. the string is not ending with only 6 characters, so grepl
will return FALSE for the OP's pattern. Instead it can be,
with(prachData, prachData[grepl("^[A-Z][A-Z][0-9]{4}", WCEL.name), ])
To show a reproducible example
v1 <- factor(c(0, 0, 'CE0001U21B2', 'CE0001U21A3',
'CE0001U21C1', 'CE0001U21B1'))
grepl("[A-Z]{2}[0-9]{4}$", v1)
#[1] FALSE FALSE FALSE FALSE FALSE FALSE
returns all FALSE
So, when we subset 'v1' based on the above index,
v1[grepl("[A-Z]{2}[0-9]{4}$", v1)]
#factor(0)
#Levels: 0 CE0001U21A3 CE0001U21B1 CE0001U21B2 CE0001U21C1
it returns a length of 0.
while
grepl("[A-Z]{2}[0-9]{4}", v1)
#[1] FALSE FALSE TRUE TRUE TRUE TRUE
Upvotes: 4