How to remove multiple pipes in data frame?

Question

Here is simple example

library(dplyr)
library(stringr)

    > dataf <- data_frame(text = c('this is a pip||e |' ,
+                              'this is |||'))
> dataf
# A tibble: 2 x 1
                text
               <chr>
1 this is a pip||e |
2        this is |||

I want to replace all the pipes in the data with an empty string. Basically I want them to disappear. However, I am only able to get rid of one of them at a time:

     > dataf %>% mutate(text = str_replace(text, '\\|+', ""))
# A tibble: 2 x 1
              text
             <chr>
1 this is a pipe |
2         this is

What is wrong here? Thanks!

in base R, `gsub("\\|", "", 'this is |||')` works as does `sub("\\|+", "", 'this is |||')`. — lmo, Jun 13 '17 at 18:01
Just confirmed that `str_replace(text, '\\|+', "")` will work. — lmo, Jun 13 '17 at 18:03
https://stackoverflow.com/questions/4736/learning-regular-expressions — jogo, Jun 13 '17 at 18:03
a ha! the gsub solution does not work with the updated example. — ℕʘʘḆḽḘ, Jun 13 '17 at 18:07
`gsub` will work. maybe you meant the `sub` solution. That regex is designed to remove the first (set of) adjacent pipes. `gsub`, on the other hand will remove them all. You could even use `gsub("|", "", 'this is a pip||e |', fixed=TRUE)` to make it a bit more readable. — lmo, Jun 13 '17 at 18:12

Psidom · Accepted Answer · 2017-06-13T18:12:40.453

3

You can use str_replace_all from stringr to remove all the matched patterns:

dataf %>% mutate(text = str_replace_all(text, '\\|', ""))

# A tibble: 2 × 1
#            text
#           <chr>
#1 this is a pipe
#2       this is

edited Jun 13 '17 at 18:12

answered Jun 13 '17 at 18:03

Psidom

209,562
33
339
356

1

I prefer Imo's answer. Let's embrace regex. – M-- Jun 13 '17 at 18:05
its nice to have multiple solution to simple problems – ℕʘʘḆḽḘ Jun 13 '17 at 18:06
2

@Masoud Just like to point out there is a difference between `str_replace_all(, "\\|",..)` and `str_replace(,"\\+",..)`; Their difference is the same as the difference between `gsub` and `sub`; You can test this string with the two options "ab||cd|". – Psidom Jun 13 '17 at 18:09
1

@Psidom I agree. That was a personal preference for this specific problem. – M-- Jun 13 '17 at 18:14

How to remove multiple pipes in data frame?

1 Answers1