Transform at the row level not column level during a dplyr chain

Question

Here is a 1 column data frame:

x <- c("   [CAD] TRAFF-S", "[CAD] SUSP-A   ", "    [CAD] CRASH", "[CAD] TRAFF-S ", "[CAD] PARKING")
x <- data.frame(x)
> x
                 x
1    [CAD] TRAFF-S
2  [CAD] SUSP-A   
3      [CAD] CRASH
4   [CAD] TRAFF-S 
5    [CAD] PARKING

I want to strip the whitespace and was looking at Trim from the qdap package.

The problem is that the input for Trim() and other functions from qdap I want to use expect a vector not a data frame. Normally that would be fine expect due to the pre-processing the vector length might change and thus re-attaching my pre-processed vector to the original df will be hard since the lengths will differ.

In the example data there's some whitespace.

Tried:

> dat.p <- x %>% Trim
> dat.p
[1] "c(2, 4, 1, 5, 3)" # I don't understand what this is doing

What is an appropriate way to iterate over each row in x within a dplyr chain? Or put another way is there a "dplyr esque" way to do this? I would like the variable x to end up being of the same format, a 1 column df witht he same number of rows and columns at the end of the transformation.

Answer 1

You have to mutate the column x , and you can use the timws function, from base R:

library(dplyr)

dat.p <- x %>% 
  mutate(x = trimws(x, 'both'))

dat.p
#               x
# 1 [CAD] TRAFF-S
# 2  [CAD] SUSP-A
# 3   [CAD] CRASH
# 4 [CAD] TRAFF-S
# 5 [CAD] PARKING