[SOLVED] Using grep with count_if (EXPSS package in R)

Using grep with count_if (EXPSS package in R)

I'm trying to count instances where a certain string appears in a dataframe (this will be a sub-string, i.e. "blue" will appear within a larger block of text), and then summarize those counts by another field.

Here's the code:

    totals_by_county <- county_data %>%
      group_by(county_data$county)%>%
      summarise(number_occurences = count(grepl('blue', county_data$color,ignore.case = TRUE))) 
    totals_by_county

And I get this error:

no applicable method for 'summarise_' applied to an object of class "logical"

Is there a way to do this in the method I'm trying to use above?

Solution

With grepl:

totals_by_county <- county_data %>%
    group_by(county) %>%
    summarise(number_occurences = sum(grepl('blue', color, ignore.case = TRUE)))

or, with count_if from expss:

totals_by_county <- county_data %>%
    group_by(county) %>%
    summarise(number_occurences = count_if(perl('blue', ignore.case = TRUE), color))

UPDATE with reproducible example:

library(dplyr)
library(expss)

county_data = data.frame(
    county = c("A", "A", "A", "B", "B"),
    color = c("blue-blue", "red", "orange-blue", "yellow", "green"),
    stringsAsFactors = FALSE)


county_data %>%
    group_by(county) %>%
    summarise(number_occurences = count_if(perl('blue', ignore.case = TRUE), color)) 

# A tibble: 2 x 2
# county number_occurences
# <chr>              <int>
# 1 A                  2
# 2 B                  0