`merge_df()` unexpected behaviour about sjmisc HOT 5 CLOSED

In essence, what I wanted to achieve was a wrapper for merge(x1, x2, all = TRUE)
but it seems like merge_df() is doing something a little different:

x1 <- efc %>% dplyr::select(1:5) %>% slice(1:10)
x2 <- efc %>% dplyr::select(3:7) %>% slice(1:10)
merge(x1, x2, all = TRUE) %>% dim()
# [1] 10  7
merge_df(x1, x2) %>% dim()
# [1] 20  7

from sjmisc.

Ok, this was a wrong description in the docs of merge_df(). I think I had an example where base merge() did the same as merge_df(), but maybe I was wrong or using a very specific example, which yielded the same results in both merge() and merge_df().

merge_df() intends to "append" two data frames, thus it is an additional feature to dplyr's join_*() functions. So, it's working as intended, just the description with merge() was wrong. I fixed the docs.

from sjmisc.

btw, dplyr's join-function preserve all label attributes, so you can use these if you need other join-operations.

from sjmisc.

Thanks for the clarifications! Yes I used dplyr::left_join() to much avail.

from sjmisc.

A typical use case is for example, if you have different country data sets that share a common amount of variables, plus some country specific variables for each country. You can then "merge" these data sets, i.e. all are "row bound", while the specific variables that only appear in one of the data sets is also added as new column, with NA values for those subsets that don't have this variable.

from sjmisc.