This function joins multiple tibbles and generates unique columns to record non-NA values when there are identical column names.
Arguments
- tibble_names
A vector of tibbles' names in your R environment.
- join_type
It should be one of 'full_join', 'left_join', 'inner_join'.
- by_col
The column that you would like join on. For example the ID column.
Details
The motivation to create this function is that when Bolin extract SNAC-N variables, he finds that for wave 3, there are overlapping information from different sources.
For example, for physician variables, there are 3 files:
All wave 3 participants.
Only cohort 1's follow-up at wave 3.
Only cohort 2's baseline.
There are overlapping of both participants and variables.
In addition, for the same participant and same variable, some are NA in one file whereas not NA in the other file.
If only using a join function, the common columns will be separated to '.x' and '.y'. And we have to use coalesce() to pick one without NA. And then we delete the '.x' and '.y' columns.
To avoid repetitive work, Bolin is writing this function to take in the information from different data files.
Examples
if (FALSE) {
## fake_snacn_ph_wave3 and fake_snacn_ph_fu contains same variable 'ph121'.
## but for some obs, in one file in NA, in the other file is not NA. E.g. Lopnr = 10
## let's just combine the info together and only keep the one without NA.
get_unique_join(tibble_names = c("fake_snacn_ph_wave3", "fake_snacn_ph_fu"),
join_type = "full_join",
by_col = "Lopnr")
}