R Finding the maximum value in subsets of observations without using summarise or filter

Question

everyone,

I have the following dataframe example, with the ID of patients (1 and 2), their category X (YES or NO), and the values of a parameter

df <- data.frame (ID = c (1, 1, 1, 1, 2, 2, 2),
                  X = c ("YES", "YES", "NO", "NO", "YES", "NO", "NO"),
                  Value = c (10, 15, 12, 13, 18, 16, 17))

df

This provides the following table:

  ID   X Value
1  1 YES    10
2  1 YES    15
3  1  NO    12
4  1  NO    13
5  2 YES    18
6  2  NO    16
7  2  NO    17

I would like to get a new column result that would give the maximum value, per patient, responding to "YES" in the column X as follows

  ID   X Value Result
1  1 YES    10     15
2  1 YES    15     15
3  1  NO    12     15
4  1  NO    13     15
5  2 YES    18     18
6  2  NO    16     18
7  2  NO    19     18

I know that I can use group_by and summarise to obtain the values, but I would like to use mutate so that I can follow all the variables that I build for this project, and for the same reason, avoid the filter function.

The following solution provides me the result column, but again I would like only one value per ID.

df %>%
  group_by(ID,X)%>%
  mutate (Result = max(Value))

     ID X     Value Result
  <dbl> <chr> <dbl>  <dbl>
1     1 YES      10     15
2     1 YES      15     15
3     1 NO       12     13
4     1 NO       13     13
5     2 YES      18     18
6     2 NO       16     19
7     2 NO       19     19

Thank you very much for your help

Answer 1

Using data.table

library(data.table)
setDT(df)[, result := max(Value[X == "YES"], na.rm = TRUE), ID]

Answer 2

What about this?

> transform(df, Result = ave(Value, ID, X, FUN = max))
  ID   X Value Result
1  1 YES    10     15
2  1 YES    15     15
3  1  NO    12     13
4  1  NO    13     13
5  2 YES    18     18
6  2  NO    16     19
7  2  NO    19     19