问题
I have two lists of dataframes that I want to combine, I have tried the following:
A <- data.frame(ID = c(1,2,3),
var1 = c(3,4,5),
var2 = c(6,3,2))
B <- data.frame(ID = c(1,2,3),
var1 = c(4,4,5),
var2 = c(6,7,3))
C <- data.frame(ID = c(1,2,3),
var1 = c(1,4,8),
var2 = c(9,2,3))
list1 <- list(A = A, B = B, C = C)
list2 <- list(A = A, B = B)
combined <- map2(list1, list2, full_join, by = 'ID')
This returns an error that the two lists are of different lengths. The only other way I have thought of is to add a blank dataframe to the second list so they are the same length.
Is it possible combine the two lists so that I get a single list where A1 has been joined with A2, B1 with B2, and C1 remains as it is?
Edit: Its been highlighted that I havent named the elements of the list, I have named them now
回答1:
If we have named lists, then we could:
list1 <- list(A = A, B = B, C = C)
list2 <- list(A = A, B = B)
x12 <- intersect(names(list1), names(list2))
x1 <- setdiff(names(list1), names(list2))
x2 <- setdiff(names(list2), names(list1))
combined <- c(
map2(list1[ x12 ], list2[ x12 ], full_join, by = 'ID'),
list1[ x1 ],
list2[ x2 ])
回答2:
If you want to keep working with positions, you can also pass placeholders and handle them in the combining logic. Something like that:
skip <- tibble(ID = integer(0))
list1 <- list(A, B, C)
list2 <- list(A, B, skip)
list3 <- list(A, skip, C)
combined <- map2(list1, list2, full_join, by = 'ID')
combined2 <- map2(list1, list3, full_join, by = 'ID')
Note how list2
and list3
would look exactly the same if not for the placeholders!
More generically:
list1 <- list(A, B, C)
list2 <- list(A, B, NULL)
list3 <- list(A, NULL, C)
combine <- function(list1, list2) {
purrr::map2(list1, list2, function(df1, df2) {
if (is.null(df1)) return(df2)
if (is.null(df2)) return(df1)
full_join(df1, df2, by = 'ID')
})
}
combined <- combine(list1, list2)
combined2 <- combine(list1, list3)
来源:https://stackoverflow.com/questions/58934147/joining-dataframes-from-lists-of-unequal-length