Regex with Positive lookhead still splits string in wrong place using strsplit()

Question

I´m trying to split a character vector containing messages right in front of a date-time indicator.

I was thinking about using strsplit() with a regular expression and perl = TRUE

Here´s some example data:

TEST <- c("05.10.17, 09:26 - Person One: How about we chill on sunday\n05.10.17, 09:27 - Person One: I could bring some beer\n05.10.17, 09:27 - Person Two: Sounds good, we could go to Lindas Party afterwards\n05.10.17, 09:27 - Person One: shit man, not LiNDA -.-\n05.10.17, 09:27 - Person Two: ???\n05.10.17, 09:28 - Person Two: You guys have history?\n05.10.17, 09:28 - Person One: She killed my family and sold their ears as souvenirs\n")

This is what I tried so far:

Cut <- unlist(strsplit(TEST,"(?=[0-3][0-9][.][0-9]{2}[.][0-9]{2}[,][ ][0-9]{2}:[0-9]{2})", perl = TRUE))
Cut

according to this website, the regex should cut the string right in front of the date-time indicator. However, the result I get looks like this, with the first character getting cut off:

 [1] "0"                                                                                   
 [2] "5.10.17, 09:26 - Person One: How about we chill on sunday\n"                         
 [3] "0"                                                                                   
 [4] "5.10.17, 09:27 - Person One: I could bring some beer\n"                              
 [5] "0"                                                                                   
 [6] "5.10.17, 09:27 - Person Two: Sounds good, we could go to Lindas Party afterwards\n"  
 [7] "0"                                                                                   
 [8] "5.10.17, 09:27 - Person One: shit man, not LiNDA -.-\n"                              
 [9] "0"                                                                                   
[10] "5.10.17, 09:27 - Person Two: ???"                                                                   
[11] "0"                                                                                   
[12] "5.10.17, 09:28 - Person Two: You guys have history?\n"                               
[13] "0"                                                                                   
[14] "5.10.17, 09:28 - Person One: She killed my family and sold their ears as souvenirs\n"

This is what the result should look like:

 [1] "05.10.17, 09:26 - Person One: How about we chill on sunday\n"                                                                                   
 [2] "05.10.17, 09:27 - Person One: I could bring some beer\n"                         
 [3] "05.10.17, 09:27 - Person Two: Sounds good, we could go to Lindas Party afterwards\n"                                                                                   
 [4] "05.10.17, 09:27 - Person One: shit man, not LiNDA -.-\n"                              
 [5] "05.10.17, 09:27 - Person Two: ???\n"                                                                                   
 [6] "05.10.17, 09:28 - Person Two: You guys have history?\n"  
 [7] 05.10.17, 09:28 - Person One: She killed my family and sold their ears as souvenirs\n"

Note: I can´t split the data at the newline indicator because some of the messages contain one or more of those in the middle of the message.

@ysth, They meant [tag:pcre] – ikegami Jan 26 '18 at 18:34 — ikegami, Jan 26 '18 at 18:34

Onyambu · Accepted Answer · 2018-01-28T04:06:23.507

You just need to create a splitting pattern when \n is followed by the date.

 strsplit(gsub("(.*?\\n)(\\d+[.]\\d+[.]\\d+)","\\1SPLITHERE\\2",TEST),"SPLITHERE")
[[1]]
[1] "05.10.17, 09:26 - Person One: How about we chill on sunday\n"                         
[2] "05.10.17, 09:27 - Person One: I could bring some beer\n"                              
[3] "05.10.17, 09:27 - Person Two: Sounds good, we could go to Lindas Party afterwards\n"  
[4] "05.10.17, 09:27 - Person One: shit man, not LiNDA -.-\n"                              
[5] "05.10.17, 09:27 - Person Two: ???\n"                                                  
[6] "05.10.17, 09:28 - Person Two: You guys have history?\n"                               
[7] "05.10.17, 09:28 - Person One: She killed my family and sold their ears as souvenirs\n"

You can also use rematches from base r

 regmatches(TEST,gregexpr(".*?\\n",TEST))
[[1]]
[1] "05.10.17, 09:26 - Person One: How about we chill on sunday\n"                         
[2] "05.10.17, 09:27 - Person One: I could bring some beer\n"                              
[3] "05.10.17, 09:27 - Person Two: Sounds good, we could go to Lindas Party afterwards\n"  
[4] "05.10.17, 09:27 - Person One: shit man, not LiNDA -.-\n"                              
[5] "05.10.17, 09:27 - Person Two: ???\n"                                                  
[6] "05.10.17, 09:28 - Person Two: You guys have history?\n"                               
[7] "05.10.17, 09:28 - Person One: She killed my family and sold their ears as souvenirs\n"

This works well in the example but could you explain what the Regex does? It looks like its somehow based on the \n string and not the actual date-time string? I guess this would cause problems if there´s a newline in the middle of the message? — Ju Ko, Jan 27 '18 at 17:00
@JuKo Ohh i see what you mean. I have added a solution that is more convenient to use — Onyambu, Jan 28 '18 at 04:05

score 1 · Answer 2 · answered Jan 26 '18 at 15:53

You can add a white character class \\s before your positive lookahead.

I have slightly changed your example to make it match your question more precisely (ie add \n inside the titles)

> TEST <- c("05.10.17, 09:26 - Person One: How about\n we chill on sunday\n05.10.17, 09:27 - Person One: I could bring some beer\n05.10.17, 09:27 - Person Two: Sounds good, we could go to Lindas Party afterwards\n05.10.17, 09:27 - Person One: shit man, not LiNDA -.-\n05.10.17, 09:27 - Person Two: ???\n05.10.17, 09:28 - Person Two: You guys have history?\n05.10.17, 09:28 - Person One: She killed my family and sold their ears as souvenirs\n")
> unlist(strsplit(TEST,"\\s(?=[0-3][0-9][.][0-9]{2}[.][0-9]{2}[,][ ][0-9]{2}:[0-9]{2})", perl = TRUE))

## [1] "05.10.17, 09:26 - Person One: How about\n we chill on sunday"                         
## [2] "05.10.17, 09:27 - Person One: I could bring some beer"                                
## [3] "05.10.17, 09:27 - Person Two: Sounds good, we could go to Lindas Party afterwards"    
## [4] "05.10.17, 09:27 - Person One: shit man, not LiNDA -.-"                                
## [5] "05.10.17, 09:27 - Person Two: ???"                                                    
## [6] "05.10.17, 09:28 - Person Two: You guys have history?"                                 
## [7] "05.10.17, 09:28 - Person One: She killed my family and sold their ears as souvenirs\n"

This does not give the desired output. Look at the output. there is `\n` at the end — Onyambu, Jan 26 '18 at 15:54
You are right ! Your solution adding a spliting pattern is better ! — Gilles San Martin, Jan 28 '18 at 17:54

score 1 · Answer 3 · answered Jan 26 '18 at 16:04

1

strsplit(TEST, '(?<=\\\n|^)(0)',perl=T)[[1]][2:7]

answered Jan 26 '18 at 16:04

Shenglin Chen

4,504
11
11

Regex with Positive lookhead still splits string in wrong place using strsplit()

3 Answers3