R data.table - alternative to lapply?

Question

I am working on a sales dataset and I need to restructure a data.table that looks like:

ID sales month
23 657   01
67 889   01
.
.
.
23 238   12

to the following format:

ID 01  02  03  ... 12
23 657 445 223 ... 238
67 889 990 765 ... 555
.
.
.

The code I currently have supplies a list of all unique IDs to lapply and the function gets all rows with the matching ID, creates a data table with one row, and joins it to the main restructured table using rbindlist.

This works for a small number of unique IDs (~600), but when the number of IDs is large, it's very slow and not feasible. Is there a more efficient way to do this? (A data.table specific approach?)

You are trying to reshape the data. Check this `data.table` answer. https://stackoverflow.com/a/36253517 — Ronak Shah, Oct 09 '20 at 01:48

score 1 · Accepted Answer · answered Oct 09 '20 at 01:43

1

We can use dcast from data.table

library(data.table)
dcast(setDT(Df1), ID ~ month, value.var = 'sales')

answered Oct 09 '20 at 01:43

akrun

874,273
37
540
662

1

dcast is great, it restructured ~70k IDs instantly. Thank you for the answer. – davey_j Oct 09 '20 at 01:57

R data.table - alternative to lapply?

1 Answers1