How do I optimize this row-wise paste in R?

Asked Dec 12 '19 at 00:45

Active Dec 12 '19 at 00:45

Viewed 128 times

I have a data frame with 1.6M rows and almost 400 columns. I'm trying to create a key for every row by concatenating the column values:

paste(df[, c(15:400)], collapse = "_")

This is taking forever, and I'm wondering if there is a faster way to perform this action? All the columns are just 0 or 1 values and the key would look something like "0_0_1_0_1_1_0_1...".

asked Dec 12 '19 at 00:45

iamsoconfused12321

2

There are lot of options here https://stackoverflow.com/questions/14568662/paste-multiple-columns-together did you try them ? – Ronak Shah Dec 12 '19 at 00:51
Do you want to paste each row's values together to get one value per row? – divibisan Dec 12 '19 at 00:55
`apply(mtcars[,3:5], 1, paste, collapse="_")` – r2evans Dec 12 '19 at 00:56
4

can i safely say that your real problem is to create a key for every row based on 386 columns? – chinsoon12 Dec 12 '19 at 01:08
Seconding this^. If your key contains all of your data, then it's hardly a key. – Daniel V Dec 12 '19 at 01:09
`do.call(paste,c(sep="-",df[15:400]))` – Onyambu Dec 12 '19 at 01:13
Why not just number the rows sequentially? Can you explain your goal that is motivating the idea of copying all the data again into one column? – Jon Spring Dec 12 '19 at 04:58

How do I optimize this row-wise paste in R?

0 Answers0