Rdata.table更新通过参考加入，但更新正确的表

Question

Context：data.table vignette中令人惊叹的新

juins的最后一部分说明了如何使用Merge sintax（x[i,on=.(id),...]）通过参考来更新左表（也在like this so so询问中进行了描述）。这是一个简化的示例：

x = data.table(id = 1:5, newvar1=LETTERS[1:5], newvar2=5:1)
               #In practice x would have more vars: newvar2, ..., newvarN
 i = data.table(id = 1:7, var1 = c('bla','ble','bli','blo','blu','blA','blS') )

#Updating the LEFT table by reference
x[i,on = .(id),
  j = `:=`(id=id,var1=var1,newvar1=newvar1)]

#Result: the column var1 (from i) is added to x by reference
print(x)
> print(x)
id newvar1   var1
<int> <char> <char>
1:     1      A    bla
2:     2      B    ble
3:     3      C    bli
4:     4      D    blo
5:     5      E    blu

我需要相反：通过参考更新

i

x = data.table(id = 1:5, newvar1=LETTERS[1:5])
               #In practice x would have more vars: newvar2, ..., newvarN 
i = data.table(id = 1:7,var1 = c('bla','ble','bli','blo','blu','blA','blS'))

#Right join then overwrite the content of i
i <- x[i,on = .(id)]

有一种方法可以通过参考来做到这一点？

在我的应用程序中（我假设为几个DT重型用户），

i

是主要数据集，拥有数百万个观察值和数十列。也就是说，我想继续添加新变量的数据集，例如来自

newvar1

，

newvar2

，...，

newvarN

。这样，我需要保留

x

的列和行（基数），而不是

i

。 Afaik的aforwriting

x

涉及制作

i

的副本，该副本会导致内存尖峰（〜双倍），该副本可能超过可用的RAM和crasterr.

我目前的解决方法是：

i

update1：我在data.table的github

上添加了一个相关问题
    

这项工作应该吗？

i_id <- i[,.(id)]     # create an auxiliary data.table with just the 'id' var
i_id <- x[i_id,on=id] #the still making the copy, but with just the 'id' columns + columns from d (instead of hundreds of columns from i) 
i[,newvar1:=i_id$newvar1]
i[,newvar2:=i_id$newvar2]
...
i[,newvarN:=i_id$newvarN]   #This works, but the code is quite large (if N is large), verbose and error prone

使用

cols <- c("newvar1", "newvar2") # or setdiff(names(x), "id")
i[, (cols) := x[.SD, on = "id", .SD, .SDcols = cols]]
# or to avoid the double .SD.
# i[, (cols) := x[i, on = "id", .SD, .SDcols = cols]]
# i[, (cols) := x[.SD, on = "id", mget(cols)]] 

i
#       id   var1 newvar1 newvar2
#    <int> <char>  <char>  <char>
# 1:     1    bla       A       F
# 2:     2    ble       B       G
# 3:     3    bli       C       H
# 4:     4    blo       D       I
# 5:     5    blu       E       J
# 6:     6    blA    <NA>    <NA>
# 7:     7    blS    <NA>    <NA>

Answer 1