正则表达式：提取部分url并在r中创建新列

Question

I have records of URL and I want to extract part of it and create new columns. In my example, I would like to consider the number after "groups" as group_id and the number of dicussion_topics as discussion_id df like:

 user  url
    1      "https://test.com/groups/3276/discussion_topics/3939"
    2      "https://test.com/groups/34/discussion_topics/11"
    3      "https://test.com/groups/3276"
    4      "https://test.com/groups/other"

I want result like

user  group_id  dicussion_id  
1      3276       3939
2      34         11
3      3276       NA
4      NA         NA

How can I do it with the regular expression in R? thx

Answer 1

dat$group_id=as.numeric(sub(".*/groups/(\\d+).*|.*","\\1",dat$url))
dat$discussion=as.numeric(sub(".*/discussion_topics/(\\d+).*|.*","\\1",dat$url))
dat
  user                                                 url group_id discussion
1    1 https://test.com/groups/3276/discussion_topics/3939     3276       3939
2    2     https://test.com/groups/34/discussion_topics/11       34         11
3    3                        https://test.com/groups/3276     3276         NA
4    4                       https://test.com/groups/other       NA         NA

Answer 2

Another version with stringi package and lookbehind regex

Update: Admittedly, the function of @Onyambu is faster. See the benchmark. Update2: Added the third version the benchmark. No improvement concerning speed.

library(stringi)
extract_info = function(x) {
  x$group = stri_extract_all_regex(x$url, "(?<=groups/)\\d+")
  x$topic = stri_extract_all_regex(x$url, "(?<=discussion_topics/)\\d+")
  x
}
extract_info(dat)
#    user                                                 url group topic
# 1    1 https://test.com/groups/3276/discussion_topics/3939  3276  3939
# 2    2     https://test.com/groups/34/discussion_topics/11    34    11
# 3    3                        https://test.com/groups/3276  3276    NA
# 4    4                       https://test.com/groups/other    NA    NA

extract_info2 = function(dat) {
dat$group_id=as.numeric(sub(".*/groups/(\\d+).*|.*","\\1",dat$url))
dat$discussion=as.numeric(sub(".*/discussion_topics/(\\d+).*|.*","\\1",dat$url))
dat
}

extract_info3 = function(data) {
  df$group_id <- as.numeric(regmatches(df$url, gregexpr(".*groups/*\\K.\\d+", df$url, perl=TRUE)))
  df$discussion <- as.numeric(regmatches(df$url, gregexpr(".*topics/*\\K.\\d+", df$url, perl=TRUE)))
  df
}

microbenchmark::microbenchmark(
  extract_info(dat)
  ,extract_info2(dat)
  ,extract_info3(dat)
)
# Unit: microseconds
#            expr     min      lq     mean   median       uq      max neval
# extract_info(dat)  152.769 160.269 172.1629 170.5325 176.0590  300.011   100
# extract_info2(dat)  99.872 106.386 120.9876 117.2415 125.7285  226.981   100
# extract_info3(dat) 285.799 301.984 378.7235 308.8925 323.3000 6684.297   100

Answer 3

Here is another option:

df$group_id <- as.numeric(regmatches(df$url, gregexpr(".*groups/*\\K.\\d+", df$url, perl=TRUE)))
df$discussion <- as.numeric(regmatches(df$url, gregexpr(".*topics/*\\K.\\d+", df$url, perl=TRUE)))