Binding rows, but not wanting to add columns where there's a mismatch

[vounder]

When using bind_rows I often want one dataset to take precedent, and to ensure those columns are kept and no more added.

On that basis it would be good to have a feature similar to left_join, right_join etc where we have top_join, bottom_join (and only matching columns too maybe?) for binding rows together.

Is there already something that does this?

[nirgrahamuk]

How about select only the names of your main df from your other dfs and bind the results. You could wrap that up in a function for yourself if you like it.

[gabriel.de.wit]

Exactly what I was thinking

library(tidyverse)

df_1 <- data.frame(x = c(1, 1, 1), y = c("a", "a", "a"))
df_2 <- data.frame(x = c(2, 2, 2), z = c("b", "b", "b"))

top_joined <- df_1 %>%
  bind_rows(df_2) %>%
  select(names(df_1))

bottom_joined <- df_1 %>%
  bind_rows(df_2) %>%
  select(names(df_2))

[vounder]

Thanks for this solution. It's not a problem doing that, it's just an extra step always.

I have just raised it as it feels like an incompleteness to the binding functions in dplyr.