The most basic function of RegEx work just like any other search.

Here I search in the text of our class survey to see who included the word “favorite” in their answer regarding their favorite work of art.

library(stringr)

str_detect(survey$fav_art, "favorite")

 [1]    NA FALSE  TRUE  TRUE FALSE  TRUE  TRUE    NA    NA  TRUE  TRUE FALSE
[13] FALSE  TRUE    NA  TRUE FALSE FALSE FALSE  TRUE FALSE FALSE  TRUE    NA

Survey Response #3:

Often, your searches will need to be more complex.

The building blocks of RegEx searches are individual characters.

This lets you do handy things like search for all numbers in a blob of text and remove them.

Select every digit in the text blob.

regexplain::view_regex(sherlock, "\\d+\\. ")

You can specify specific parameters for the text you want to search for, like length.

Here I select all 4 digit numbers, which I assume are years.

You can perform a similar trick for phone numbers or other structured text.

Select all 4 digit numbers.

regexplain::view_regex(zelda, "\\d{4}")

Sometimes you won’t know exactly what you want to find.

In this case, you can use “wildcards,” which let you say things like: “give me everything after this word until the end of a sentence.”

You always have to be careful though, otherwise it might just select everything!

Use some wildcards to find any email address in the text.

regexplain::view_regex(email,
  "([[:alnum:]_.-]+)@([[:alnum:].-]+)\\.([[:alpha:].]{2,6})")

Sometimes you will want to get data out of large blobs of unstructured text.

Using the building blocks we’ve covered so far, you can specify “groups” you want to capture.

Take this blob of text from the spring schedule at Smith.

[1] "Thursday, January 26\nClasses begin at 8 a.m.\nWednesday, February 1\nLast day to add a course online\nWednesday, February 8\nLast day to drop a course online; last day to add a Five College course\nWednesday, February 15\nLast day to add a Smith course\nThursday, February 23: Rally Day\nRally Day is highlighted by an all-college convocation at which distinguished alumnae are awarded the Smith College Medal. Afternoon classes are canceled.\nSaturday, March 11–Sunday, March 19\nSpring recess. Houses close at 10 a.m. on Saturday, March 11, and reopen at 1 p.m. on Sunday, March 19.\nMonday, April 3–Friday, April 14\nAdvising and course registration for the fall 2023 semester"

stringr::str_detect() will give you a TRUE or FALSE if any part of a string matches the pattern you provide.

It will work on a vector, returning a single TRUE or FALSE per element.

test_vec = c("I like turtles",
             "I like cats",
             "I don't like raptors")

stringr::str_detect(test_vec, "like")

[1] TRUE TRUE TRUE

Case matters! (by default)

stringr::str_detect(test_vec, "Like")

[1] FALSE FALSE FALSE

stringr::str_count() will return a number for each element in a vector, telling you how many times a match was found.

This could be useful for counting dates, key words, or anything else you can imagine.

For example, consider searching through a record of newspapers for how often a specific politician appears.

stringr::str_extract_all() will extract your matches from the vector elements.

The result will be a list of length X, matching how many vectors you have it.

The elements of that list will be vectors of length Y containing all the matches.

Topic

Lab 4

To-Do

• Finish Worksheet