counting specific words across multiple columns in

问题:

I have a data frame like this

df <- data.frame(id=c(1, 2, 3, 4, 5), staple_1=c("potato", "cassava","rice","fruit","coffee"), staple_2=c("cassava","beer","peanuts","rice","yams"), staple_3=c("rice","peanuts","fruit","fruit","rice"))

I also have a character vector like this

staples<-c("potato","cassava","rice","yams")

I would like to create a new variable that is the row sum of the occurrence of any of the words in the "staples" character vector. The outcome of which should look like this

df$staples<-c(3,1,1,1,2)

I have tried several approaches and nothing has worked so far. My actual dataframe is much bigger and character vector has 20 or more words in it. I'm sure there is a simple solution but that I am missing it somehow.

回答1:

A simple apply will do it.

apply(df, 1, function(x) sum(staples %in% x))
#[1] 3 1 1 1 2

df$staples <- apply(df, 1, function(x) sum(staples %in% x))

回答2:

Here is a tidyverse solution.

library(tidyverse)

df <- data_frame(id = c(1, 2, 3, 4, 5), 
             staple_1 = c("potato", "cassava", "rice", "fruit", "coffee"), 
             staple_2 = c("cassava", "beer", "peanuts", "rice", "yams"), 
             staple_3 = c("rice", "peanuts", "fruit", "fruit", "rice"))

staples_vect <- c("potato", "cassava", "rice", "yams")

df %>% 
     mutate(staples = pmap_int(select(., starts_with("staple_")), ~sum(c(...) %in% staples_vect)))

It seems best to avoid using the same name as for both a global variable and a column in a dataframe when using dplyr.

回答3:

Another approach is to loop through the columns, use %in%, to create a list of logical vectors and add (+) with Reduce

Reduce(`+`, lapply(df[-1], `%in%`, staples))
#[1] 3 1 1 1 2

---

Or with rowSums by converting the dataset into matrix (without the 'id' column), use %in% to convert to logical vector, assign the dim to convert to same dimensions as the dimensions of df[-1] and get the rowSums

rowSums(`dim<-`(as.matrix(df[-1]) %in% staples, dim(df[-1])))
#[1] 3 1 1 1 2