How to chunk multiple user_id together to form another column in a data.table?-CodePudding

I need to group and sort my primary key (user_id) together by chunking them together. So, I need to introduce a new column that sort of act like a counter. Essentially, I need to make something like this:

chunk_user |   user    |     item     |
1          |  200401   |    78832     |
1          |  200401   |    95718     |
1          |  200401   |    24161     |
2          |  200402   |    12437     |
2          |  200402   |    61490     |
2          |  200402   |    45956     |

from something like this:

user id | isbn | rating
    123 | 4567 | 2
    129 | 7890 | 3
    127 | 4450 | 0
    123 | 9972 | 1

I tried using the group_by() to start but not only it doesn't work like the way I wanted, it doesn't even sort my data properly.

table_rated <- tbl_ratings %>% group_by(user_id) %>% arrange(isbn, ratings, sort = TRUE)

I really appreciate any help!

CodePudding user response：

It's not entirely clear, but this may be close to what you're looking for.

You can use rleid from data.table to enumerate different user_ids.

When using arrange you can include .by_group to sort first by the grouping variable.

Let me know if this is what you had in mind.

library(tidyverse)
library(data.table)

tbl_ratings %>%
  group_by(user_id) %>%
  arrange(isbn, rating, .by_group = T) %>%
  ungroup %>%
  mutate(chunk_user = rleid(user_id))

Output

  user_id  isbn rating chunk_user
    <dbl> <dbl>  <int>      <int>
1     123  4567      2          1
2     123  9972      1          1
3     127  4450      0          2
4     129  7890      3          3