dplyr find grouped minimum when column equals value-CodePudding

I have a df that looks as follow:

GROUP    DATE    VALUE    TIMEFRAME
  A     1/1/20    1           0
  A     1/2/20    0           0
  A     1/3/20   .5           1
  A     1/4/20   .3           1
  A     1/5/20    1           0
  B     1/3/20   .1           1
  B     1/4/20   .2           1
  B     1/5/20    1           0

What I want to do is calculate the minimum per group but only when TIMEFRAME == 1 so the output would be as follows:

GROUP    DATE    VALUE    TIMEFRAME   MIN
  A     1/1/20    1           0        .3
  A     1/2/20    0           0        .3
  A     1/3/20   .5           1        .3
  A     1/4/20   .3           1        .3
  A     1/5/20    1           0        .3
  B     1/3/20   .1           1        .1
  B     1/4/20   .2           1        .1
  B     1/5/20    1           0        .1

I have tried using filter() here but I need to keep all of the rows

CodePudding user response：

Within the mutate, subset the 'VALUE' based on the logical expression created with TIMEFRAME, get the min on those subset of values after grouping by 'GROUP'

library(dplyr)
df1 %>%
  group_by(GROUP) %>% 
  mutate(MIN = min(VALUE[TIMEFRAME == 1])) %>%
  ungroup

-output

# A tibble: 8 × 5
  GROUP DATE   VALUE TIMEFRAME   MIN
  <chr> <chr>  <dbl>     <int> <dbl>
1 A     1/1/20   1           0   0.3
2 A     1/2/20   0           0   0.3
3 A     1/3/20   0.5         1   0.3
4 A     1/4/20   0.3         1   0.3
5 A     1/5/20   1           0   0.3
6 B     1/3/20   0.1         1   0.1
7 B     1/4/20   0.2         1   0.1
8 B     1/5/20   1           0   0.1

data

df1 <- structure(list(GROUP = c("A", "A", "A", "A", "A", "B", "B", "B"
), DATE = c("1/1/20", "1/2/20", "1/3/20", "1/4/20", "1/5/20", 
"1/3/20", "1/4/20", "1/5/20"), VALUE = c(1, 0, 0.5, 0.3, 1, 0.1, 
0.2, 1), TIMEFRAME = c(0L, 0L, 1L, 1L, 0L, 1L, 1L, 0L)), 
class = "data.frame", row.names = c(NA, 
-8L))