I currently have a list of vectors where each number in the list represents the count of a different word in the file.
I would like to change this list to be a dataframe where the rownames are the filenames, and columns are the words (sorted alphabetically with only one column per word), and each observation the count of a certain word, where all words used in any file is included (i.e. if file a includes a word that file b does not include, then the count of the word in file b is 0).
So essentially the current code right now looks like:
file1 <- c(1,5,7,2)
names(file1) <- c("a", "by", "her", "the")
file2 <- c(10,5,2)
names(file2) <- c("a", "and", "to")
list(file1, file2)
What I would like is:
df <- data.frame(matrix(nrow=2, ncol=6, byrow=T, data=c(1, 0, 5, 7, 2, 0,
10, 5, 0,0,0,2)))
colnames(df) <- c("a", "and", "by", "her", "the", "to")
rownames(df) <- c("file1", "file2")
df
Thanks.
The fill argument of rbindlist function from data.table package can come in handy here.
library(data.table)
nm = c("file1", "file2")
d = rbindlist(lapply(mget(nm), function(x) data.frame(t(x))), fill = TRUE)
d = as.data.frame(d)
row.names(d) = nm
d
# a by her the and to
#file1 1 5 7 2 NA NA
#file2 10 NA NA NA 5 2
To reorder d and replace NA with 0, further steps are necessary
d = d[,order(colnames(d))]
d = replace(d, is.na(d), 0)