CODEWITHSUNDEEP
rfunctional-programmingclosures
Magnetar
Magnetar

Reputation: 197

Don't understand Cannot Coerce type 'closure' Error

I see this is a common issue but I can't understand what to do from reading other posts or trying to understand functional programming which is new to me. Functions are closures in R, encapsulating the environment they were created in? The code I have is:

# Remove numbers from text
minus_TextNum <- function(df, new.df){
  new.df <- mutate(df, text = gsub(x = text, pattern = "[0-9]+|\\(.*\\)", replacement = "")) %>%  # and/or whatever's in brackets
    unnest_tokens(input = text, output = word) %>% 
    filter(!word %in% c(stop_words$word, "patient")) %>% 
    group_by(id) %>% 
    summarise(text = paste(word, collapse = " "))
  return(new.df)
}

minus_TextNum(TidySymptoms)

Error is as follows:

Error: Problem with mutate() column text. ℹ text = gsub(x = text, pattern = "[0-9]+|\\(.*\\)", replacement = ""). x cannot coerce type 'closure' to vector of type 'character'

I don't understand what type closure is, and this is a simple function that works on a simple dataset I created to test. Problem arises when I use the real-world dataset.

Any feedback appreciated. Reproducible sample below:

# Remove numbers and/or anything in brackets

# Test Data
mydata <- data.frame(id = 1:8,
                     text = c("112773 Nissan Micra, Car, (10 pcs)",
                              "112774 Nissan Micra, Car, (10 pcs)",
                              "112775 Nissan Micra, Car, (10 pcs)",
                              "112776 Volkswagon Beetle, Car, (3 pcs)",
                              "112777 Toyota Corolla, Car, (12 pcs)",
                              "112778 Nissan Micra, Car, (10 pcs)",
                              "112779 Toyota Prius, Car, (9 pcs)",
                              "112780 Toyota Corolla, Car, (12 pcs)"),
                     stringsAsFactors = F)

library(dplyr)
library(tidytext)

# remove numbers from text data
data(stop_words)
minus_TextNum <- function(df, new.df){
  new.df <- mutate(df, text = gsub(x = text, pattern = "[0-9]+|\\(.*\\)", replacement = "")) %>%  # and/or whatevers in brackets
    unnest_tokens(input = text, output = word) %>% 
    filter(!word %in% c(stop_words$word, "car")) %>% 
    group_by(id) %>% 
    summarise(text = paste(word, collapse = " "))
  return(new.df)
}


minus_TextNum(mydata)

dput(head(TidySymptoms, n = 10)) structure(list(word = c("epiglottis", "swelled", "hinder", "swallowing", "pictures", "benadryl", "tylenol", "approximately", "30", "min" )), row.names = c(NA, 10L), class = "data.frame")

Upvotes: 0

Views: 824

Answers (1)

Ronak Shah
Ronak Shah

Reputation: 388982

TidySymptoms data has no id column in it. Assuming it's a mistake and you have that already in your data you can do the following changes in the function.

  • There is no need to pass df.new to the function.
  • The column in TidySymptoms is called as word but you are using text in the function.

Try this code.

minus_TextNum <- function(df){

  df.new <- mutate(df, text = gsub(x = word, pattern = "[0-9]+|\\(.*\\)", replacement = "")) %>%  
    unnest_tokens(input = text, output = word) %>% 
    filter(!word %in% c(stop_words$word, "patient")) %>% 
    group_by(id) %>% 
    summarise(text = paste(word, collapse = " "))
    return(new.df)
}

minus_TextNum(TidySymptoms)

Upvotes: 1

Related Questions