r - 使用将连续变量分成两组的值构建表

Question

我正在寻找你的帮助。我试图将一个连续变量分为两组，我把这个例子放在我想要做的事情上:

x=data.frame(v1=c(1,1,2,2,3,4,5,6,9,9,11,2,4,45,67,89,1,1,5,5,5,6,6,6,9,9,9,11,11,8,8,8,51,90,40,15,30,11,8,9,9,1,5,5,100,67,78,98,34,25))

我尝试将我的连续变量分成两组，初始值为 20，然后:

g1=data.frame(x[x$v1>20,])
g2=data.frame(x[x$v1<=20,])

在我计算 g1 和 g2 的均值和 sd 之后:

mean(g1$x.x.v1...20...)=62.61538
mean(g2$x.x.v1....20...)=6.216216
sd(g1$x.x.v1...20...)=26.80963
sd(g2$x.x.v1....20...)=3.55227
length(g1$x.x.v1...20...)= 13
length(g2$x.x.v1....20...)=37

在此之后，我想要一个显示如下内容的表格:

Value   Mean.G1  SD.G1  Mean.G2  SD.G2  N.G1 N.G2
20        62.61    26.8   6.21     3.55  13   37

但这张表不仅适用于 20 我想为具有不同值的向量构建该表，例如一个具有 10 个元素的向量，从 20 开始，并在步骤 op 20 中增加，这样的向量 v=c(10,30,50,70,90,110,130,150,170,190) .
我等这个问题清楚了。谢谢

Answer 1

我会使用 reshape2 和 plyr，

library(plyr) ; library(reshape2)
v=c(10,20,30,50,70,90,110,130,150,170,190) # added 20 for checking
# create new dichotomy id variable
l1 = llply(v, function(.v) transform(x, test = x[["v1"]] <= .v))
names(l1) = v # name list elements for later reference
all = melt(l1, id=c("v1","test")) # merge data.frames together
# summarise the data by groups
results = ddply(all, c("L1","test"), summarise, 
          mean = mean(v1), sd=sd(v1), length=length(v1))

导致

arrange(results, as.numeric(L1))

    L1  test      mean        sd length
1   10 FALSE 48.500000 32.505656     18
2   10  TRUE  5.343750  2.902828     32
3   20 FALSE 62.615385 26.809633     13
4   20  TRUE  6.216216  3.552270     37
5   30 FALSE 69.000000 23.870484     11
6   30  TRUE  7.307692  5.907862     39
7   50 FALSE 80.000000 17.270950      8
8   50  TRUE  9.619048 10.245647     42
9   70 FALSE 91.000000  8.717798      5
10  70  TRUE 13.088889 16.555447     45
11  90 FALSE 99.000000  1.414214      2
12  90  TRUE 17.625000 23.951747     48
13 110  TRUE 20.880000 28.456655     50
14 130  TRUE 20.880000 28.456655     50
15 150  TRUE 20.880000 28.456655     50
16 170  TRUE 20.880000 28.456655     50
17 190  TRUE 20.880000 28.456655     50