使用完全连接来匹配两个不同数据帧中的值

Question

I have two dataframes.The first data frame consists of four columns 1) ID, 2) Site, 3) Depth, and 4) Density. The second dataframe consists of 1) ID 2) Site, and 3) Choice.

df1

  ID Sites Depth Density
  1     B   0.2       0
  2     B   0.2       1
  3     D   0.3       0
  4     D   0.3       1
  5     B   0.2       2

df2

  ID Sites Choice 
  1     A    No
  1     B    Yes     
  1     C    No
  1     D    No
  2     A    No
  2     B    Yes
  2     C    No
  2     D    No
  3     A    No
  3     B    No
  3     C    No
  3     D    Yes
  4     A    No
  4     B    No
  4     C    No
  4     D    Yes
  5     A    No
  5     B    Yes
  5     C    No
  5     D    No

What I am trying to do is add a column to df2 that has the densities in each site when the ID has a "Yes". Below is what I want the output to be:

Desired Output

  ID Sites Choice Depth  Density
  1     A    No     0.1     0
  1     B    Yes    0.2     0 
  1     C    No     0.3     0 
  1     D    No     0.4     0
  2     A    No     0.1     0
  2     B    Yes    0.2     1
  2     C    No     0.3     0
  2     D    No     0.4     0
  3     A    No     0.1     0
  3     B    No     0.2     1
  3     C    No     0.3     0
  3     D    Yes    0.4     0
  4     A    No     0.1     0
  4     B    No     0.2     1
  4     C    No     0.3     0
  4     D    Yes    0.4     1
  5     A    No     0.1     0
  5     B    Yes    0.2     2
  5     C    No     0.3     0
  5     D    No     0.4     1

I've tried using the following but it doesn't work:

     df3<-df2 %>%
     full_join(df1, by = c("ID", "Sites")) %>%
     group_by(ID) %>%
     mutate(Density = Density[Choice == "Yes"]) %>%
     distinct(ID, Sites, .keep_all = TRUE) 

Thank you for your help, stackoverflow community.

Answer 1

Is this what you are looking for?

df1<-data.frame(stringsAsFactors=FALSE,
                ID = c(1, 2, 3, 4, 5),
                Sites = c("B", "B", "D", "D", "B"),
                Depth = c(0.2, 0.2, 0.3, 0.3, 0.2),
                Density = c(0, 1, 0, 1, 2)
)

df2<-data.frame(stringsAsFactors=FALSE,
                ID = c(1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5),
                Sites = c("A", "B", "C", "D", "A", "B", "C", "D", "A", "B", "C", "D",
                          "A", "B", "C", "D", "A", "B", "C", "D"),
                Choice = c("No", "Yes", "No", "No", "No", "Yes", "No", "No", "No", "No",
                           "No", "Yes", "No", "No", "No", "Yes", "No", "Yes", "No",
                           "No")
)

df3<-df2 %>%
  left_join(df1, by = c("ID", "Sites")) %>%
  mutate(Density=if_else(Choice=="Yes",Density, 0))

Answer 2

Data.

  df1 <- data.frame(
      ID = seq(from = 1, to = 5, by = 1),
      Sites = c("B", "B", "D", "D", "B"),
      Depth = c(0.2, 0.2, 0.3, 0.3, 0.2),
      Density = c(0,1,0,1,2)
    )

qw <- c("No", "Yes")
  
df2 <- data.frame(
  ID = c(rep(1:5, times = 5)),
  Sites = sample(LETTERS[1:4], size = 25, replace = TRUE),
  Choice = sample(qw, size = 25, replace = T)
)

But I am slightly confused by what you need from this reprex. Your statement makes it seem that you want a subsetted DF3, which only contains records where CHOICE == "YES" is TRUE, and ID = ID, and Site = Site. Obviously this will result in very few records.

df3 <- merge(df1, df2, by = c("Sites", "ID")) %>% 
filter(Choice == "Yes") # only one record from the dummy data.

If the Choice 'Yes' is a property of the ID, but not Site, here is an option.

df3 <- df1 %>% 
select(-Sites) %>% 
  left_join(df2 %>% filter(Choice == "Yes"), by = "ID")

I guess I am just confused whether both Sites and ID's need to match. If so their are just not many records here.

Answer 3

A data.table solution:

df1[df2[ID %in% df2[Choice == "Yes", unique(ID)], ], 
    on = .(ID, Sites)][is.na(Density), 
                           Density := 0][]

What is in there:

• df2[Choice == "Yes", unique(ID) filters where choice is yes and returns unique IDs. We'll need them to filter df2 .
• df2[ID %in% df2[Choice == "Yes", unique(ID)], ] filters only the cases where ID matches the IDs where there is at least one "yes" in choices.
• df1[df_x, on =.(ID, Sites)] makes a left join of df1 with whatever is in df_x (in our case, the filtered df2 .
• [is.na(Density), Density:= 0] filters the rows where Density is NA and only in those rows assigns 0 to Density.
• [] Prints to screen the resuling data, you don't really need it.