Extract strings between characters

Question

Here's a substring I have

> substring(reut2.000[4,], regexpr(">",reut2.000[3,]) + 1)
[1] "<D>el-salvador</D><D>usa</D><D>uruguay</D></PLACES>"

And I would like to extract all the characteres between <D> and </D> .
In this case, the output will be

"el-salvador","use","uruguay"

So far I have tried

gsub(".*<D>\\s*|</D>.*", "", tmp)

where tmp is the substring, and it returns "uruguay" .

How do I modify it so that it returns all the places?

Answer 1

You have an XML file (<<== is likely exactly the file you have, too). Note that said link is to an example file from the tmparallel package and there are many places in that package that have code that works with it .

Work with XML as XML. Do not regex it.

xdf$places in the following snippet has what you are looking for but since this is likely a file being used in class on text mining, you may eventually need all the other bits extracted into the data frame.

  library(xml2)
  library(tidyverse)

  download.file(
    "https://raw.githubusercontent.com/noahhl/tmparallel/master/pkg/inst/texts/reuters-21578.xml",
    "~/Data/reuters-21578.xml"
  )

  reut <- read_xml("~/Data/reuters-21578.xml")

  xml_find_all(reut, "//REUTERS") %>%
    map_df(~{

      xml_attrs(.x) %>%
        as.list() %>%
        as_data_frame() -> xdf

      xdf$date <- xml_find_first(.x, ".//DATE") %>% xml_text(trim=TRUE)

      #### NOTE THAT THIS FOLLOWING LINE IS THE DATA YOU ASKED FOR IN THE EXAMPLE

      xdf$places <- list(xml_find_all(.x, ".//PLACES/D") %>% xml_text(trim=TRUE))

      xdf$people <- list(xml_find_all(.x, ".//PEOPLE/D") %>% xml_text(trim=TRUE))

      xdf$orgs <- list(xml_find_all(.x, ".//ORGS/D") %>% xml_text(trim=TRUE))

      xdf$exchanges <- list(xml_find_all(.x, ".//EXCHANGES/D") %>% xml_text(trim=TRUE))

      xdf$companies <- list(xml_find_all(.x, ".//COMPANIES/D") %>% xml_text(trim=TRUE))

      xdf$uknown <- xml_find_first(.x, ".//UNKNOWN") %>% xml_text(trim=TRUE)

      xdf$text_title <- xml_find_first(.x, ".//TEXT/TITLE") %>% xml_text(trim=TRUE)

      xdf$text_dateline <- xml_find_first(.x, ".//TEXT/DATELINE") %>% xml_text(trim=TRUE)

      xdf$text_body <- xml_find_first(.x, ".//TEXT/BODY") %>% xml_text(trim=TRUE)

      xdf

    }) -> text_df

Output:

  text_df
  ## # A tibble: 10 x 15
  ##    TOPICS LEWISSPLIT CGISPLIT  OLDID NEWID date      places  people  orgs 
  ##    <chr>  <chr>      <chr>     <chr> <chr> <chr>     <list>  <list>  <lis>
  ##  1 YES    TRAIN      TRAINING… 5544  1     26-FEB-1… <chr [… <chr [… <chr…
  ##  2 NO     TRAIN      TRAINING… 5545  2     26-FEB-1… <chr [… <chr [… <chr…
  ##  3 NO     TRAIN      TRAINING… 5546  3     26-FEB-1… <chr [… <chr [… <chr…
  ##  4 NO     TRAIN      TRAINING… 5547  4     26-FEB-1… <chr [… <chr [… <chr…
  ##  5 YES    TRAIN      TRAINING… 5548  5     26-FEB-1… <chr [… <chr [… <chr…
  ##  6 YES    TRAIN      TRAINING… 5549  6     26-FEB-1… <chr [… <chr [… <chr…
  ##  7 NO     TRAIN      TRAINING… 5550  7     26-FEB-1… <chr [… <chr [… <chr…
  ##  8 YES    TRAIN      TRAINING… 5551  8     26-FEB-1… <chr [… <chr [… <chr…
  ##  9 YES    TRAIN      TRAINING… 5552  9     26-FEB-1… <chr [… <chr [… <chr…
  ## 10 YES    TRAIN      TRAINING… 5553  10    26-FEB-1… <chr [… <chr [… <chr…
  ## # ... with 6 more variables: exchanges <list>, companies <list>,
  ## #   uknown <chr>, text_title <chr>, text_dateline <chr>, text_body <chr>

  glimpse(text_df)
  ## Observations: 10
  ## Variables: 15
  ## $ TOPICS        <chr> "YES", "NO", "NO", "NO", "YES", "YES", "NO", "YE...
  ## $ LEWISSPLIT    <chr> "TRAIN", "TRAIN", "TRAIN", "TRAIN", "TRAIN", "TR...
  ## $ CGISPLIT      <chr> "TRAINING-SET", "TRAINING-SET", "TRAINING-SET", ...
  ## $ OLDID         <chr> "5544", "5545", "5546", "5547", "5548", "5549", ...
  ## $ NEWID         <chr> "1", "2", "3", "4", "5", "6", "7", "8", "9", "10"
  ## $ date          <chr> "26-FEB-1987 15:01:01.79", "26-FEB-1987 15:02:20...
  ## $ places        <list> [<"el-salvador", "usa", "uruguay">, "usa", "usa...
  ## $ people        <list> [<>, <>, <>, <>, <>, <>, <>, <>, <>, <>]
  ## $ orgs          <list> [<>, <>, <>, <>, <>, <>, <>, <>, <>, <>]
  ## $ exchanges     <list> [<>, <>, <>, <>, <>, <>, <>, <>, <>, <>]
  ## $ companies     <list> [<>, <>, <>, <>, <>, <>, <>, <>, <>, <>]
  ## $ uknown        <chr> "C T\nf0704reute\nu f BC-BAHIA-COCOA-REVIEW   02...
  ## $ text_title    <chr> "BAHIA COCOA REVIEW", "STANDARD OIL <SRD> TO FOR...
  ## $ text_dateline <chr> "SALVADOR, Feb 26 -", "CLEVELAND, Feb 26 -", "HO...
  ## $ text_body     <chr> "Showers continued throughout the week in\nthe B...

  str(head(text_df, 2))
  ## Classes 'tbl_df', 'tbl' and 'data.frame':    2 obs. of  15 variables:
  ##  $ TOPICS       : chr  "YES" "NO"
  ##  $ LEWISSPLIT   : chr  "TRAIN" "TRAIN"
  ##  $ CGISPLIT     : chr  "TRAINING-SET" "TRAINING-SET"
  ##  $ OLDID        : chr  "5544" "5545"
  ##  $ NEWID        : chr  "1" "2"
  ##  $ date         : chr  "26-FEB-1987 15:01:01.79" "26-FEB-1987 15:02:20.00"
  ##  $ places       :List of 2
  ##   ..$ : chr  "el-salvador" "usa" "uruguay"
  ##   ..$ : chr "usa"
  ##  $ people       :List of 2
  ##   ..$ : chr 
  ##   ..$ : chr 
  ##  $ orgs         :List of 2
  ##   ..$ : chr 
  ##   ..$ : chr 
  ##  $ exchanges    :List of 2
  ##   ..$ : chr 
  ##   ..$ : chr 
  ##  $ companies    :List of 2
  ##   ..$ : chr 
  ##   ..$ : chr 
  ##  $ uknown       : chr  "C T\nf0704reute\nu f BC-BAHIA-COCOA-REVIEW   02-26 0105" "F Y\nf0708reute\nd f BC-STANDARD-OIL-<SRD>-TO   02-26 0082"
  ##  $ text_title   : chr  "BAHIA COCOA REVIEW" "STANDARD OIL <SRD> TO FORM FINANCIAL UNIT"
  ##  $ text_dateline: chr  "SALVADOR, Feb 26 -" "CLEVELAND, Feb 26 -"
  ##  $ text_body    : chr  "Showers continued throughout the week in\nthe Bahia cocoa zone, alleviating the drought since early\nJanuary an"| __truncated__ "Standard Oil Co and BP North America\nInc said they plan to form a venture to manage the money market\nborrowin"| __truncated__

Answer 2

Here is one option using grepexpr and regmatches to capture all matches in your text:

input <- c("<D>el-salvador</D><D>usa</D><D>uruguay</D></PLACES>")
m <- gregexpr("(?<=<D>).*?(?=</D>)", input, perl=TRUE)
regmatches(input, m)[[1]]

[1] "el-salvador" "usa"         "uruguay"

Demo

Note that it is generally not advisable to use regex to parse HTML/XML or similar content. One reason for this is that there could be nested tags, causing a simple regex to break.

Answer 3

Another option using gsub:

temp <- "<D>el-salvador</D><D>usa</D><D>uruguay</D></PLACES>"

temp <- unlist(strsplit(gsub("<D>|</D>|</PLACES>", " ", x = temp ), split = " "))
temp <- temp[temp != ""]