R find min and max for each group based on other row

Question

I am stuck in transforming an existing data frame in R, using dplyr (but open for other options. I am running out of ideas and nothing brings me closer to the required result. The data frame looks like that:

data.frame("group" = c('a', 'a', 'b', 'b', 'c', 'c' ), "condition"= c(1, 2, 1, 2, 1,2 ), "X1" = c(2010,'x', 2011, 'x', 2010, 'x'), "X2" = c(2011,'x', 2012, 'x', 2011, 'x'), "X3" = c(2012,'x', 2013, 'x', 2012, 'x'), "X4" = c(2013,'x', 2014, 'x', 2013, 'x'), "X5" = c(2014,'', 2015, 'x', 2014, 'x'), "X6" = c(2015,'', 2015, '', 2015, ''))

For each group, the new data frame should show the earliest and last year (condition 1) that contains an 'x' in condition 2.

The result should look like:

 data.frame("group" = c('a', 'b',  'c'  ), "min"= c(2010, 2011, 2010), "max" = c(2013, 2015, 2014))

Answer 1

With tidyverse you can try the following approach. First, put your data into long form targeting your year columns. Then, group_by both group and name (which contains the year) and only include subgroups that have a value of x , and keep rows that have condition of 1. Then group_by just group and summarise to get the min and max years. Note, you may wish to convert your year data to numeric after removing x by filtering on condition .

library(tidyverse)

df1 %>%
  pivot_longer(cols = -c(group, condition)) %>%
  group_by(group, name) %>%
  filter(any(value == "x"), condition == 1) %>%
  group_by(group) %>%
  summarise(min = min(value),
            max = max(value))

Output

# A tibble: 3 x 3
  group min   max  
  <chr> <chr> <chr>
1 a     2010  2013 
2 b     2011  2015 
3 c     2010  2014

Answer 2

in Base-R

results <- df[df$condition==1,1:2]
results <- cbind(results, t(apply(df[df$condition==1,3:ncol(df)],1,function(x)c(Min=min(x),Max=max(x)))))



  group condition  Min  Max
1     a         1 2010 2015
3     b         1 2011 2015
5     c         1 2010 2015