R - Group by variable and then assign a unique ID [duplicate]

Question

I am interested in de-identifying a sensitive data set with both time-fixed and time-variant values. I want to (a) group all cases by social security number, (b) assign those cases a unique ID and then (c) remove the social security number.

Here's an example data set:

personal_id    gender  temperature 111-11-1111      M        99.6 999-999-999      F        98.2 111-11-1111      M        97.8 999-999-999      F        98.3 888-88-8888      F        99.0 111-11-1111      M        98.9 

Any solutions would be very much appreciated.

Answer 1

dplyr has a group_indices function for creating unique group IDs

library(dplyr) data <- data.frame(personal_id = c("111-111-111", "999-999-999", "222-222-222", "111-111-111"),                        gender = c("M", "F", "M", "M"),                        temperature = c(99.6, 98.2, 97.8, 95.5))  data$group_id <- data %>% group_indices(personal_id)  data <- data %>% select(-personal_id)  data   gender temperature group_id 1      M        99.6        1 2      F        98.2        3 3      M        97.8        2 4      M        95.5        1 

Or within the same pipeline (https://github.com/tidyverse/dplyr/issues/2160):

data %>%      mutate(group_id = group_indices(., personal_id)) 

Answer 2

dplyr::group_indices() is deprecated as of dplyr 1.0.0. dplyr::cur_group_id() should be used instead:

df %>%  group_by(personal_id) %>%  mutate(group_id = cur_group_id())    personal_id gender temperature group_id   <chr>       <chr>        <dbl>    <int> 1 111-11-1111 M             99.6        1 2 999-999-999 F             98.2        3 3 111-11-1111 M             97.8        1 4 999-999-999 F             98.3        3 5 888-88-8888 F             99          2 6 111-11-1111 M             98.9        1