CODEWITHSUNDEEP
rjoinmergedplyr
Blundering Ecologist
Blundering Ecologist

Reputation: 1315

Replacing multiple observations from one column with values from another column in R

I am trying to replace values from two columns with values from another two columns. This is a rather basic question and has been asked by python users, however I am using R.

I have a df that looks like this (only on a much larger scale [>20,000]):

squirrel_id    locx    locy    dist
6391           17.5    10.0    50.0
6391           17.5    10.0    20.0
6391           17.5    10.0    15.5
8443           20.5    1.0     800
6025           -5.0    -0.5    0.0

I need to, for 63 squirrels, replace their locx and locy values.

I normally replace values with the following code:

library(dplyr)    

df <- df %>%
   mutate(locx = ifelse (squirrel_id=="6391", "12.5", locx),
         locy = ifelse (squirrel_id=="6391", "15.5", locy),
         locx = ifelse (squirrel_id=="8443", "2.5", locx),
         locy = ifelse (squirrel_id=="8443", "80", locy)) #etc for 63 squirrels

Which would give me:

squirrel_id    locx    locy    dist
6391           12.5    10.0    50.0
6391           12.5    10.0    20.0
6391           12.5    10.0    15.5
8443           2.5     80.0    800
6025           -5.0    -0.5    0.0

But this is creating an extra 126 lines of code and I suspect there is a simpler way to do this.

I do have all the new locx and locy values in a separate df, but I do not know how to join the two dataframes by squirrel_id without it messing up the data.

df with the values that need to replace the ones in the old df:

squirrel_id    new_locx    new_locy   
6391           12.5        15.5 
8443           2.5         80
6025           -55.0       0.0

How can I do this more efficiently?

Upvotes: 0

Views: 332

Answers (2)

Cole
Cole

Reputation: 11255

Using @ANG's data, here's a data.table solution. It joins and updates the original df by reference. 

library(data.table)

setDT(df)
setDT(df2)

df[df2, on = c('squirrel_id'), `:=` (locx = new_locx, locy = new_locy) ]

df

   squirrel_id  locx locy  dist
1:        6391  12.5 15.5  50.0
2:        6391  12.5 15.5  20.0
3:        6391  12.5 15.5  15.5
4:        8443   2.5 80.0 800.0
5:        6025 -55.0  0.0   0.0
6:        5000  18.5 12.5  10.0

See also:

how to use merge() to update a table in R

Replace a subset of a data frame with dplyr join operations

R: Updating a data frame with another data frame

Upvotes: 0

nghauran
nghauran

Reputation: 6768

You can left_join the two data frames and then use an if_else statement to get the right locx and locy. Try out:

library(dplyr)
df %>% left_join(df2, by = "squirrel_id") %>%
        mutate(locx = if_else(is.na(new_locx), locx, new_locx), # as suggested by @echasnovski, we can also use locx = coalesce(new_locx, locx)
               locy = if_else(is.na(new_locy), locy, new_locy)) %>% # or locy = coalesce(new_locy, locy)
        select(-new_locx, -new_locy)
# output
  squirrel_id  locx locy  dist
1        6391  12.5 15.5  50.0
2        6391  12.5 15.5  20.0
3        6391  12.5 15.5  15.5
4        8443   2.5 80.0 800.0
5        6025 -55.0  0.0   0.0
6        5000  18.5 18.5  10.0 # squirrel_id 5000 was created for an example of id 
# present if df but not in df2

Data

df <- structure(list(squirrel_id = c(6391L, 6391L, 6391L, 8443L, 6025L, 
5000L), locx = c(17.5, 17.5, 17.5, 20.5, -5, 18.5), locy = c(10, 
10, 10, 1, -0.5, 12.5), dist = c(50, 20, 15.5, 800, 0, 10)), class = "data.frame", row.names = c(NA, 
-6L))
df2 <- structure(list(squirrel_id = c(6391L, 8443L, 6025L), new_locx = c(12.5, 
2.5, -55), new_locy = c(15.5, 80, 0)), class = "data.frame", row.names = c(NA, 
-3L))

Upvotes: 1

Related Questions