R how to create a dictionary of unique values [duplicate] - r

Question

This question already has answers here:
Split character column into several binary (0/1) columns
(7 answers)
Closed 2 years ago.
I have a column in a dataframe that contains multiple values like this
fruits
1 apple,banana
2 banana,peaches
3 peaches
4 mango
Is there a way to create a dictionary of unique values for fruits which
is will create a new column fruits with values :
fruits = apple,banana,peaches,mango
UPDATE: I need the value as a column and not a list of just unique values . So that I can create a final dataframe that would have the following :
fruits fruit_apple fruit_banana fruit_mango fruit_peacheas
1 apple,banana 1 0 0 0
2 banana,peaches 0 1 0 1
3 peaches 0 0 0 1
4 mango 0 0 1 0

Answer 1

We can do this easily with cSplit_e from splitstackshape
library(splitstackshape)
cSplit_e(df1, "fruits", ",", type = "character", fill = 0)
# fruits fruits_apple fruits_banana fruits_mango fruits_peaches
#1 apple,banana 1 1 0 0
#2 banana,peaches 0 1 0 1
#3 peaches 0 0 0 1
#4 mango 0 0 1 0
data
df1 <- structure(list(fruits = c("apple,banana", "banana,peaches", "peaches",
"mango")), .Names = "fruits", class = "data.frame", row.names = c("1",
"2", "3", "4"))

Answer 2

Do you want the new column to be that concatenated list repeated? Sorry, it's not particularly clear. Assuming that's the case though, and that your data.frame consists of strings not factors;
df <- read.delim(
text="fruits
apple,banana
banana,peaches
peaches
mango",
sep="\n",
header=TRUE,
stringsAsFactors=FALSE)
df
#> fruits
#> 1 apple,banana
#> 2 banana,peaches
#> 3 peaches
#> 4 mango
df$uniquefruits <- paste0(unique(unlist(strsplit(df$fruits, split=","))), collapse=",")
df
#> fruits uniquefruits
#> 1 apple,banana apple,banana,peaches,mango
#> 2 banana,peaches apple,banana,peaches,mango
#> 3 peaches apple,banana,peaches,mango
#> 4 mango apple,banana,peaches,mango
Or do you mean taking only the values from your first fruits column that are not duplicated elsewhere?
Update: Based on comments, I think this is what you're after:
uniquefruits <- unique(unlist(strsplit(df$fruits, split=",")))
uniquefruits
#> [1] "apple" "banana" "peaches" "mango"
df2 <- cbind(df,
sapply(uniquefruits,
function(y) apply(df, 1,
function(x) as.integer(y %in% unlist(strsplit(x, split=","))))))
df2
#> fruits apple banana peaches mango
#> 1 apple,banana 1 1 0 0
#> 2 banana,peaches 0 1 1 0
#> 3 peaches 0 0 1 0
#> 4 mango 0 0 0 1
In theory, you could do this with dplyr but I can't figure out how to automate the column processing for the rowwise mutate (anyone know how?)
library(dplyr)
df %>% rowwise() %>% mutate(apple = as.integer("apple" %in% unlist(strsplit(fruits, ","))),
banana = as.integer("banana" %in% unlist(strsplit(fruits, ","))),
peaches = as.integer("peaches" %in% unlist(strsplit(fruits, ","))),
mango = as.integer("mango" %in% unlist(strsplit(fruits, ","))))
#> Source: local data frame [4 x 5]
#> Groups: <by row>
#>
#> # A tibble: 4 x 5
#> fruits apple banana peaches mango
#> <chr> <int> <int> <int> <int>
#> 1 apple,banana 1 1 0 0
#> 2 banana,peaches 0 1 1 0
#> 3 peaches 0 0 1 0
#> 4 mango 0 0 0 1

Answer 3

with base R:
fruits <- sort(unique(unlist(strsplit(as.character(df$fruits), split=','))))
cols <- as.data.frame(matrix(rep(0, nrow(df)*length(fruits)), ncol=length(fruits)))
names(cols) <- fruits
df <- cbind.data.frame(df, cols)
df <- as.data.frame(t(apply(df, 1, function(x){fruits <- strsplit(x['fruits'], split=','); x[unlist(fruits)] <- 1;x})))
df
fruits apple banana mango peaches
1 apple,banana 1 1 0 0
2 banana,peaches 0 1 0 1
3 peaches 0 0 0 1
4 mango 0 0 1 0

Answer 4

You can use below steps,
1) Just split dataframe by comma using strsplit function.
2) Unlist a split list of vectors into a single vector.
3) Then take unique of list.fruits character vector.
Here is the solution,
# DataFrame of fruits
f <- c("apple,banana","banana,peaches","peaches","mango")
fruits <- as.data.frame(f)
# fruits dataframe
f
#1 apple,banana
#2 banana,peaches
#3 peaches
#4 mango
list.fruits <- unlist(strsplit(f,split=","))
unique.fruits <- unique(list.fruits)
# Result
unique.fruits
[1] "apple" "banana" "peaches" "mango"