I am struggling to 'translate' a regex expression from stringi/stringr to quanteda's kwic function.
How can I get all instances of "Jane Mayer", regardless of whether she has a middle name or not. Note that I don't have a list of all existing middle names in the data. So defining multiple patterns (one for each middle name) wouldn't be possible.
Many thanks!
library(quanteda)
library(tidyverse)
txt <- c("this is Jane Alexandra Mayer",
"this is Jane Mayer",
"this is Jane Eli Mayer",
"this is Jane Burger")
txt_token <- tokens(txt)
my_pattern <- c("Jane .* Mayer")
kwic(txt_token, pattern=phrase(my_pattern), valuetype = "regex")
#> Keyword-in-context with 2 matches.
#> [text1, 3:5] this is | Jane Alexandra Mayer |
#> [text3, 3:5] this is | Jane Eli Mayer |
my_pattern <- c("Jane .? Mayer")
kwic(txt_token, pattern=phrase(my_pattern), valuetype = "regex")
#> Keyword-in-context with 2 matches.
#> [text1, 3:5] this is | Jane Alexandra Mayer |
#> [text3, 3:5] this is | Jane Eli Mayer |
my_pattern <- c("Jane.* Mayer")
kwic(txt_token, pattern=phrase(my_pattern), valuetype = "regex")
#> Keyword-in-context with 1 match.
#> [text2, 3:4] this is | Jane Mayer |
my_pattern <- c("Jane . Mayer")
kwic(txt_token, pattern=phrase(my_pattern), valuetype = "regex")
#> Keyword-in-context with 2 matches.
#> [text1, 3:5] this is | Jane Alexandra Mayer |
#> [text3, 3:5] this is | Jane Eli Mayer |
With stringr I'll would simply use:
str_extract(txt, regex("Jane.* Mayer"))
#> [1] "Jane Alexandra Mayer" "Jane Mayer" "Jane Eli Mayer"
#> [4] NA
```
<sup>Created on 2021-11-28 by the [reprex package](https://reprex.tidyverse.org) (v2.0.1)</sup>
It seems you need to pass another pattern to match exactly Jane Mayer:
kwic(txt_token, pattern=phrase(c("Jane .* Mayer", "Jane Mayer")), valuetype = "regex")
# => Keyword-in-context with 3 matches.
# [text1, 3:5] this is | Jane Alexandra Mayer |
# [text2, 3:4] this is | Jane Mayer |
# [text3, 3:5] this is | Jane Eli Mayer |