In matthewhirschey/bespokelearnr: Bespoke data science learning

Using Regular Expressions

That solution worked in this case, but was not very elegant, and might not work for all cases (what if there was a 'great aunt' in the list?)

Or here is a more specific case for this data set.

How many patients have a father with a history of disease? But we don't want to include grandfathers in the results.

We can use something called Regular Expressions, aka Regex, to solve this

Using Regular Expressions {.build}

Think of regex as a separate language, with it's own code, syntax, and rules.

Regex rules allow complex matching patterns for strings, to ensure matching exactly the content desired

It is far too complex to cover in its entirely here, but here is one specific example.

GOAL: identify all of the patients that have a father with a history of disease, but excluding grandfathers in the results.

Regular Expression Example {.build}

Example

We want to start with recognizing father.

But then we want to make sure that we capture both Father and father. To accept either case f in the first spot we add (F|f), so now our regex looks like (F|f)ather

Lastly, we want this pattern to appear at the beginning of the word, so we add the regex ^ symbol.

Our completed regex looks like:

str_count(heart_joined$family_history, "^(F|f)ather")

Regex Exercise

Go to code/
Open 09_stringr.Rmd
Complete the exercise to count mothers.

Regex resources

• The stringr cheatsheet shown above is a great resource
• Here the tidyverse website with a link to its vignette

matthewhirschey/bespokelearnr documentation built on Oct. 11, 2020, 12:57 a.m.