Resolved: Filter rows by paired sites in R using limited matching characters

dat <- data.frame(var1 = c('Green Island_one_before', 'Green Island_two_before', 'Green Island_one_after', 'Pink Island_one_before', 'Pink Island_two_after', 'Pink Island_five_after'),
                 var2 = c('Green Island_three_before', 'Green Island_two_after', 'Green Island_one_after', 'Pink Island_four_before', 'Pink Island_two_before', 'Pink Island_six_after'),
                 value = c(1, 7, 4, 2, 8, 3))
[/code]

</p>



Using the example data above, it should return rows #2 and 5.<br> 



<h2>Answer:</h2>



I would start by putting each variable in a column. That is, separate <code>var1</code> and <code>var2</code> in their component parts, the site, location and time. After that, anything one needs to do will become much more straightforward.  E.g.<br> 


<p>[code]library(tidyverse)</p>
<p>dat %>%<br />
  separate(var1, into = c('site_v1', 'location_v1', 'time_v1'), sep = '_') %>%<br />
  separate(var2, into = c('site_v2', 'location_v2', 'time_v2'), sep = '_') %>%<br />
  filter(site_v1 == site_v2, location_v1 == location_v2, time_v1 != time_v2)