r - 当键列数不同时合并data.table

Question

我试图从文档中了解data.table中的逻辑，并且有点不清楚。我知道我可以尝试一下，看看会发生什么，但我想确保没有病理情况，因此想知道逻辑的实际编码方式。当两个data.table对象的键列数不同时，例如a的键数为2，而b的键数为3，并且您运行c <- a[b]，a和b会简单地合并在前两个键列上，还是会自动合并a中的第三列到b中的第三个键列？一个例子:

require(data.table)
a <- data.table(id=1:10, t=1:20, v=1:40, key=c("id", "t"))
b <- data.table(id=1:10, v2=1:20, key="id")
c <- a[b]

这应该选择与 a中的 id键列匹配的 b行。例如，对于 id==1中的 b， b中有2行， a中有4行，应该在 c中生成8行。确实确实是这样:

> head(c,10)
    id  t  v v2
 1:  1  1  1  1
 2:  1  1 21  1
 3:  1 11 11  1
 4:  1 11 31  1
 5:  1  1  1 11
 6:  1  1 21 11
 7:  1 11 11 11
 8:  1 11 31 11
 9:  2  2  2  2
10:  2  2 22  2

尝试它的另一种方法是:

d <-b[a]

这应该做同样的事情:对于 a中的每一行，都应该选择 b中的匹配行:由于 a包含一个额外的键列 t，因此该列不应用于匹配，而只能基于第一个键列进行联接， id应该完成。似乎是这种情况:

> head(d,10)
    id v2  t  v
 1:  1  1  1  1
 2:  1 11  1  1
 3:  1  1  1 21
 4:  1 11  1 21
 5:  1  1 11 11
 6:  1 11 11 11
 7:  1  1 11 31
 8:  1 11 11 31
 9:  2  2  2  2
10:  2 12  2  2

有人可以确认吗？需要明确的是:是任何合并中曾经使用的 a的第三键列，还是 data.table仅使用两个表的 min(length(key(DT)))。

Answer 1

好问题。首先，正确的术语是(来自?data.table):

[A data.table] may have one key of one or more columns. This key can be used for row indexing instead of rownames.

因此，“键”(单数)不是“键”(复数)。目前，我们可以摆脱“ key ”的束缚。但是，将来添加辅助 key 时，可能会有多个 key 。每个键(单数)可以具有多个列(复数)。

否则，你是绝对正确的。以下段落在v1.8.2中进行了改进，基于其他人也感到困惑的反馈。从 ?data.table:

When i is a data.table, x must have a key. i is joined to x using x's key and the rows in x that match are returned. An equi-join is performed between each column in i to each column in x's key; i.e., column 1 of i is matched to the 1st column of x's key, column 2 to the second, etc. The match is a binary search in compiled C in O(log n) time. If i has fewer columns than x's key then many rows of x will ordinarily match to each row of i since not all of x's key columns will be joined to (a common use case). If i has more columns than x's key, the columns of i not involved in the join are included in the result. If i also has a key, it is i's key columns that are used to match to x's key columns (column 1 of i's key is joined to column 1 of x's key, column 2 to column 2, and so on) and a binary merge of the two tables is carried out. In all joins the names of the columns are irrelevant. The columns of x's key are joined to in order, either from column 1 onwards of i when i is unkeyed, or from column 1 onwards of i's key.

在注释之后，在v1.8.3(在R-Forge上)中，现在显示为(以粗体显示的更改):

When i is a data.table, x must have a key. i is joined to x using x's key and the rows in x that match are returned. An equi-join is performed between each column in i to each column in x's key; i.e., column 1 of i is matched to the 1st column of x's key, column 2 to the second, etc. The match is a binary search in compiled C in O(log n) time. If i has fewer columns than x's key then not all of x's key columns will be joined to (a common use case) and many rows of x will (ordinarily) match to each row of i. If i has more columns than x's key, the columns of i not involved in the join are included in the result. If i also has a key, it is i's key columns that are used to match to x's key columns (column 1 of i's key is joined to column 1 of x's key, column 2 of i's key to column 2 of x's key, and so on for as long as the shorter key) and a binary merge of the two tables is carried out. In all joins the names of the columns are irrelevant; the columns of x's key are joined to in order, either from column 1 onwards of i when i is unkeyed, or from column 1 onwards of i's key. In code, the number of join columns is determined by min(length(key(x)),if (haskey(i)) length(key(i)) else ncol(i)).