multicolumn csv text into presence absence matri - r

Question

I have a csv file like this:
col1 col2 col3
r1 a,b,c e,f g
r2 h,i j,k
r3 l m,n,o
some cells have multiple text comma separated, some have single and some have none.I want to convert this like:
col1 col2 col3
a 1 0 0
b 1 0 0
c 1 0 0
e 0 1 0
f 0 1 0
g 0 0 1
h 1 0 0
i 1 0 0
j 0 0 1
k 0 0 1
l 1 0 0
m 0 1 0
n 0 1 0
o 0 1 0
Any suggestion? I tried pivot table in excel but not getting the desired output.
Thanks in advance.
Best Regards
Zillur

Answer 1

not sure whether this is the shortest solution (probably not) but it produces the desired output. Basically, we go through all three columns and count the occurrences of the strings and get a long format data frame that we then flip to the wide format you want.
library(tidyr)
library(purrr)
df <- data_frame(col1 = c("a,b,c", "h,i", "l"),
col2 = c("e,f", "", "m,n,o"),
col3 = c("g", "j,k", ""))
let_df <- map_df(df, function(col){
# map_df applies the function to each column of df
# split strings at "," and unlist to get vector of letters
letters <- unlist(str_split(col, ","))
# delete ""
letters <- letters[nchar(letters) > 0]
# count occurrences for each letter
tab <- table(letters)
# replace with 1 if occurs more often
tab[tab > 1] <- 1
# create data frame from table
df <- data_frame(letter = names(tab), count = tab)
return(df)
}, .id = "col") # id adds a column col that contains col1 - col3
# bring data frame into wide format
let_df %>%
spread(col, count, fill = 0)

Answer 2

Such a great problem to solve. Here is my take on it in base R:
col1 <- c("a,b,c","h,i","l")
col2 <- c("e,f","","m,n,o")
col3 <- c("g","j,k","")
data <- data.frame(col1, col2, col3, stringsAsFactors = F)
restructure <- function(df){
df[df==""] <- "missing"
result_rows <- as.character()
l <- list()
for (i in seq_along(colnames(df)) ){
df_col <- sort(unique(unlist(strsplit(gsub(" ", "",toString(df[[i]])), ","))))
df_col <- df_col[!df_col %in% "missing"]
result_rows <- sort(unique(c(result_rows, df_col)))
l[i] <- list(df_col)
}
result <- data.frame(result_rows)
for (j in seq_along(l)){
result$temp <- NA
result$temp[match(l[[j]], result_rows)] <- 1
colnames(result)[colnames(result)=="temp"] <- colnames(df)[j]
}
result[is.na(result)] <- 0
return(result)
}
> restructure(data)
# result_rows col1 col2 col3
#1 a 1 0 0
#2 b 1 0 0
#3 c 1 0 0
#4 e 0 1 0
#5 f 0 1 0
#6 g 0 0 1
#7 h 1 0 0
#8 i 1 0 0
#9 j 0 0 1
#10 k 0 0 1
#11 l 1 0 0
#12 m 0 1 0
#13 n 0 1 0
#14 o 0 1 0