转置数据帧 [英] Transposing a data frame

问题描述

我有一个问题，关于将数据帧重塑(如果正确的话)到它的转置版本.我想像这样:

I have a question about re-shaping (if that's the right word) a data frame to a transposed version of it. I want to take something like:

A   B   C   
1   6   1
1   18  1
1   21  1
3   18  1
3   21  1
4   6   1
4   18  1
4   20  1
4   21  1

并将其转换为如下数据框:

And turn it into a dataframe like:

A   B_1 C_1 B_2 C_2 B_3 C_3 ...
1   6   1   18  1   21  1
3   18  1   21  1   
4   6   1   18  1   20  1   21  1

R中是否有我不知道的go-to函数，还是应该遍历此过程? 谢谢！

Is there some go-to function in R that I'm unaware of or should I loop over this process? Thanks!

推荐答案

DF <- read.table(text="A   B   C   
1   6   1
1   18  1
1   21  1
3   18  1
3   21  1
4   6   1
4   18  1
4   20  1
4   21  1",header=TRUE)

首先创建一个索引变量:

First create an index variable:

library(plyr)
DF <- ddply(DF,.(A),transform,No=seq_along(A))
#   A  B C No
# 1 1  6 1  1
# 2 1 18 1  2
# 3 1 21 1  3
# 4 3 18 1  1
# 5 3 21 1  2
# 6 4  6 1  1
# 7 4 18 1  2
# 8 4 20 1  3
# 9 4 21 1  4

现在将data.frame融化并转换为宽格式:

Now melt the data.frame and cast to wide format:

library(reshape2)
DF <- melt(DF,id.vars=c("A","No"))
res <- dcast(DF,A~No+variable)
#   A 1_B 1_C 2_B 2_C 3_B 3_C 4_B 4_C
# 1 1   6   1  18   1  21   1  NA  NA
# 2 3  18   1  21   1  NA  NA  NA  NA
# 3 4   6   1  18   1  20   1  21   1

公式中的顺序很重要:

res <- dcast(DF,A~variable+No)
#   A B_1 B_2 B_3 B_4 C_1 C_2 C_3 C_4
# 1 1   6  18  21  NA   1   1   1  NA
# 2 3  18  21  NA  NA   1   1  NA  NA
# 3 4   6  18  20  21   1   1   1   1

这篇关于转置数据帧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！