在data.table中多次使用:= [英] Using := multiple times in data.table
问题描述
我经常发现自己在同一数据表上使用:=
进行了一系列的链式计算。例如,类似这样的
I frequently find myself doing a long series of chained calculations using :=
on the same data table. For example, something like this
test = data.table(1:10, 1:10, 1:10, 1:10)
test[, V1 := V1^2]
test[, V2 := V1*V2]
test[, V3 := V2/V3]
test[, V4 := sqrt(V3)]
test[, new := letters[V4]]
必须在每行1上写 test [,...]
需要更长的时间键入(这不是我可以应付的大问题) 。但是,更重要的是,还从视觉上分散了计算的流程和内容。我宁愿写这样的东西
Having to write test[, ...]
on every line 1) takes longer to type (not a big issue I can cope with that). But, more importantly, also distracts visually from the flow and content of the calculation. I would much rather write something like
test[, {
V1 := V1^2
V2 := V1*V2
V3 := V2/V3
V4 := sqrt(V3)
new := letters[V4]
}]
但这会引发的错误,您已经用{}包裹了:=,但是:=必须是{}
中唯一的东西。
But this throws an error of You have wrapped := with {} which is ok but then := must be the only thing inside {}
.
我知道我会写
within(test, {
V1 = V1^2
V2 = V1*V2
V3 = V2/V3
V4 = sqrt(V3)
new = letters[V4]
})
但这会降低使用:= $的效率c $ c>
我尝试编写一个提供此功能的函数
I tried writing a function to provide this ability
with.dt = function(dt, expressions){
e = strsplit(expressions,'\n')
for (i in 1:length(e)){
dt[, eval(parse(text = e[i]))]
}
dt
}
with.dt(test, "
V1 := V1^2;
V2 := V1*V2;
V3 := V2/V3;
V4 := sqrt(V3);
new := letters[V4];
")
但这不会更改数据的内容。表
But this does not change the content of the data.table
是否有任何语法允许 c版本内的外观整洁,同时保留<$的按引用分配属性c $ c>:=
?
Is there any syntax that permits the neat look of the within
version, while retaining the assign by reference property of :=
?
推荐答案
有两种不同的方法可以做到这一点。
There are a couple different ways to do this.
这里使用的原始测试矩阵:
Original test matrix used here:
v1 v2
1: 1 3
2: 2 4
首先,我们可以做这样的事情:
First, we can do something like this:
test[,`:=`(v1 = v1^2,
v2 = v1*v2)
输出:
v1 v2
1 3
4 8
或者,如果
test[, `:=`(v1 = v1 <- v1^2, v2 = v2 * v1)]
输出:
v1 v2
1 3
4 16
或者,我们可以类似地运行它:
Or, similarly we could run this:
test[,c("v1","v2") := list(v1^2,v1^2*v2)]
输出:
v1 v2
1 3
4 16
这篇关于在data.table中多次使用:=的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!