Created
March 8, 2018 14:16
-
-
Save tomncooper/22d94052449de164abc9cb9f26f454c5 to your computer and use it in GitHub Desktop.
Toy examle of using agrep to match food groups for participant food choices
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
library(tidyr) | |
# User data dataframe | |
user.foods = rbind("milk", "apples", "bread", "ice cream", "boxed fruits") | |
ID = rbind(1, 2, 3, 4, 5) | |
user.data = cbind(ID, user.foods) | |
colnames(user.data) = c("ID", "User.Food") | |
user.data = as.data.frame(user.data) | |
# Food Groups dataframe | |
foods = rbind("Cream", "Milk", "Yoghurt", "Bread", "Ice cream cones", | |
"Orange", "Carrots", "Milk chocolate") | |
foodcodes = rbind("Dairy", "Dairy", "Dairy", "Grains", "Dairy", | |
"Fruit and vegetables", "Fruit and vegetables", "Sweets") | |
food.groups = cbind(foods, foodcodes) | |
colnames(food.groups) = c("Food", "Food.Code") | |
food.groups = as.data.frame(food.groups) | |
# Find row indexes from food.groups that match pattern | |
match.food = function(pattern){ | |
matches = agrep(pattern, food.groups$Food, ignore.case = TRUE, | |
value = FALSE) | |
return(matches) | |
} | |
# Add matched indexes to user data | |
user.data$matched.indexes = sapply(user.data$User.Food, FUN=match.food) | |
# Convert the no matches (which is a list with zero in) to NA | |
check = function(x) { | |
if(sum(x) == 0) { | |
return(TRUE) | |
} else { | |
return(FALSE) | |
} | |
} | |
user.data[sapply(user.data$matched.indexes, check), ]$matched.indexes = NA | |
# Convert to long format | |
long.user.data = separate_rows(user.data, matched.indexes, convert=FALSE) | |
# Filter out the weird "c" and "" entries that separate_rows add because | |
# apparently no one has ever come across a situation where a cell contains an | |
# actual list and not a list in the form of a comma separated string! | |
long.user.data = long.user.data[long.user.data$matched.indexes != "c" & | |
long.user.data$matched.indexes != "" | |
| is.na(long.user.data$matched.indexes), ] | |
# Add in the food group data | |
user.food.matches = cbind(long.user.data, | |
food.groups[long.user.data$matched.indexes, | |
c("Food", "Food.Code"]) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment