加入和合并不会在 R 中返回正确的行数

Question

I have two dataframes which share a common column (named sys_loc_code). The first dataframe (df1) has 1033 rows. The second dataframe (df2) has 2751.

I would like to combine df1 and df2 so get a new dataframe with all columns found in df1 and df2 keeping only rows from df1.

I have tried join , left_join , and inner_join (from dplyr ) and a simple merge . Each of these returns 2057 rows, and I think it should only be returning 1033 to match what is in df1 . How do I return only rows from df1?

I cannot share the datasets that caused this problem. However, after a bit of consultation, I can recreate the problem with this minimal example:

df1 <-
  data.frame(
    sys_loc_code = c("A", "B", "C")
    , df1Val = 1
  )


df2 <-
  data.frame(
    sys_loc_code = c("A", "B", "B", "C", "D")
    , df2Val = c(1, 1, 2, 1, 1)
  )

left_join(df1, df2)

Returns 4 rows while df1 only has three rows.

Answer 1

The most issue is that df2$sys_loc_code contains multiple entries for some of the values in df1$sys_loc_code .

df1$sys_loc_code has only 3 values, but one of them ("B") is present twice in df2$sys_loc_code , meaning those merges will return 4 rows. eg

left_join(df1, df2)

gives

  sys_loc_code df1Val df2Val
1            A      1      1
2            B      1      1
3            B      1      2
4            C      1      1

So, the short answer to your question may be that the results actually are "correct" based on the code you are writing. If you want something different to happen (eg, only one entry from df2 per match), you will likely need to decide exactly what output you want.

For example, if you want the first entry from df2 :

left_join(
  df1
  , df2 %>%
    group_by(sys_loc_code) %>%
    slice(1)
)

gives

  sys_loc_code df1Val df2Val
1            A      1      1
2            B      1      1
3            C      1      1


left_join(
  df1
  , df2 %>%
    group_by(sys_loc_code) %>%
    summarise(df2Val = mean(df2Val))
)

gives

  sys_loc_code df1Val df2Val
1            A      1    1.0
2            B      1    1.5
3            C      1    1.0

and

left_join(
  df1
  , df2 %>%
    mutate(aVarToSortOn = 1:n()) %>%
    group_by(sys_loc_code) %>%
    slice(which.max(aVarToSortOn))
)

gives

  sys_loc_code df1Val df2Val aVarToSortOn
1            A      1      1            1
2            B      1      2            3
3            C      1      1            4

If you know you have unique values in a column, you could also use filter to select which match to keep from df2