R: Removing Whitespace + Delimiter

Question

I'm fairly new to the R language. So I have this vector containing the following:

> head(sampleVector)

[1] "| txt01 |   100 |         200 |       123.456 |           0.12345 |"
[2] "| txt02 |   300 |         400 |       789.012 |           0.06789 |"

I want to extract the lines and break each into separate pieces, with a data value per piece. I want to get a list resultListthat eventually would print out the following:

> head(resultList)`

[[1]]`  
[1] ""   "txt01"    "100"       "200"     "123.456"        "0.12345" 

[[2]]`  
[1] ""   "txt02"    "300"       "400"     "789.012"        "0.06789"

I am struggling with the strsplit() notation and I have tried and got the following code so far:

resultList  <- strsplit(sampleVector,"\\s+[|] | [|]\\s+ | [\\s+]")`          
#would give me the following output`

# [[1]]`    
# [1] "| txt01"    "100"       "200"     "123.456"        "0.12345 |"

Anyway I can get the output the one strsplit call? I am guessing my notation to distinguish the delimiter + whitespace is wrong. Any help on this would be good.

Rich Scriven · Answer 1 · 2014-10-21T01:57:20.310

4

Here's one way. This first removes the | from the vector with gsub. Then it uses strsplit on the spaces (or any number of spaces). Probably a bit easier that way.

strsplit(gsub("|", "", sampleVector, fixed=TRUE), "\\s+")
# [[1]]
# [1] ""        "txt01"   "100"     "200"     "123.456" "0.12345"
#
# [[2]]
# [1] ""        "txt02"   "300"     "400"     "789.012" "0.06789"

Here's an interesting alternative using scan that might be useful, and will probably be quite fast.

lapply(sampleVector, function(y) {
    s <- scan(text = y, what = character(), sep = "|", quiet = TRUE)
    (g <- gsub("\\s+", "", s))[-length(g)]
})
# [[1]]
# [1] ""        "txt01"   "100"     "200"     "123.456" "0.12345"
#
# [[2]]
# [1] ""        "txt02"   "300"     "400"     "789.012" "0.06789"

edited Oct 21 '14 at 01:57

answered Oct 21 '14 at 01:22

Rich Scriven

97,041
11
181
245

Wow thanks mate. I was looking into `gsub` but I didn't think of putting it within the `strsplit` function. Thanks. Will accept as answer (in around 6 minutes when SO allows me too) ahahahah thanks again for the help P.S. I wish i can upvote this ahahhaha – 12341234 Oct 21 '14 at 01:26
1

No worries. I really like these types of problems :). Also, not sure if you want to keep the first `""` element. It can be removed easily with `lapply(resultList, "[", -1)` – Rich Scriven Oct 21 '14 at 01:27
ahahahahha i'm getting the hang of it. I am unfamiliar with the notations like similar to `\\s`. Thanks again. Wish of luck for the SF Giants :) – 12341234 Oct 21 '14 at 01:29
Ah I would need it when I do my matrix manipulations later on but I'll keep that in mind. – 12341234 Oct 21 '14 at 01:29

thelatemail · Accepted Answer · 2014-10-21T01:59:41.373

4

Another strsplit option which I nearly missed:

strsplit(test,"[| ]+")
#[[1]]
#[1] ""        "txt01"   "100"     "200"     "123.456" "0.12345"
# 
#[[2]]
#[1] ""        "txt02"   "300"     "400"     "789.012" "0.06789"

...and my original answer because regmatches is my favourite function of late:

regmatches(test,gregexpr("[^| ]+",test))
#[[1]]
#[1] "txt01"   "100"     "200"     "123.456" "0.12345"
#
#[[2]]
#[1] "txt02"   "300"     "400"     "789.012" "0.06789"

To break it down as requested:

[| ]+ is a regex searching for single or repeated instances + of a space or a pipe |
[^| ]+ is a regex searching for single or repeated instances + of any character not ^ a space or a pipe |
gregexpr finds all the instances of this pattern and returns the start locations and length of the matching patterns.
regmatches extracts all the patterns from test that are matched by gregexpr

edited Oct 21 '14 at 01:59

answered Oct 21 '14 at 01:40

thelatemail

91,185
12
128
188

Oh, sure. Come with all that fancy stuff. – Rich Scriven Oct 21 '14 at 01:41
@RichardScriven - I do like to swoop in with an alternative ;-) – thelatemail Oct 21 '14 at 01:41
1

Did you see my `scan` alternative? Thinking outside the box – Rich Scriven Oct 21 '14 at 01:42
@RichardScriven - `scan` is neat - you could potentially concoct something with `read.table` too. – thelatemail Oct 21 '14 at 01:44
I think OP wants to capture the blanks as well – rawr Oct 21 '14 at 01:44
Do you care to elaborate on this interesting looking entity? :) – 12341234 Oct 21 '14 at 01:45
1

That `strsplit` solution is money in the bank. @MagnumOpus - I think you should accept this answer instead – Rich Scriven Oct 21 '14 at 01:58
Okay so any symbols in the brackets `[| ]` are the symbols we are putting in the current context. What does the `+` notation after the square brackets indicate? – 12341234 Oct 21 '14 at 01:59
@MagnumOpus - `+` represents repeats of whatever text comes before it. – thelatemail Oct 21 '14 at 02:00
Okay. I will study and look into regex notations more. @thelatemail Thanks again. – 12341234 Oct 21 '14 at 02:05

score 0 · Answer 3 · answered Oct 21 '14 at 01:33

0

May try strsplit first and the gsub:

sapply(strsplit(xx, '\\|'), function (x) gsub("^\\s+|\\s+$", "", x))
     [,1]     
[1,] ""       
[2,] "txt01"  
[3,] "100"    
[4,] "200"    
[5,] "123.456"
[6,] "0.12345"

answered Oct 21 '14 at 01:33

rnso

23,686
25
112
234

i wanted to return a list with each list component containing the split characters. but thanks anyway! – 12341234 Oct 23 '14 at 21:49

R: Removing Whitespace + Delimiter

3 Answers3