R宽到长整形并带有列名

Question

I have data in this format

A1 A2 B1 B2  C1  C2
10  5 11  5  21  10

And I want to convert it to:

  1  2
A 10 5
B 11 5
C 21 10

How can I do it in R?

Answer 1

We can gather into 'long' format, then separate the 'key' column into two by splitting before the numeric part, spread it to 'wide' and change the 'key1' column to row names

library(tidyverse)
gather(df1) %>%
    separate(key, into = c('key1', 'key2'), sep="(?=\\d)") %>% 
    spread(key2, value) %>% 
    column_to_rownames('key1')
#  1  2
#A 10  5
#B 11  5
#C 21 10

data

df1 <- structure(list(A1 = 10L, A2 = 5L, B1 = 11L, B2 = 5L, C1 = 21L, 
     C2 = 10L), class = "data.frame", row.names = c(NA, -1L))

Answer 2

The tags to the question are r, reshape and reshape2 so we show solutions using each of those.

1) xtabs A base R solution is the following.

let <- gsub("\\d", "", names(DF))
num <- gsub("\\D", "", names(DF))
tab <- xtabs(unlist(DF) ~ let + num)

giving:

> tab
   num
let  1  2
  A 10  5
  B 11  5
  C 21 10

or for a data frame:

cbind(let = rownames(tab), as.data.frame.matrix(tab))

giving:

  let  1  2
A   A 10  5
B   B 11  5
C   C 21 10

2) reshape Another base R solution is the following. let and num are from above.

varying <- split(names(DF), num)
reshape(DF, dir = "long", varying = varying, v.names = names(varying),
  times = unique(let), timevar = "let")[-4]

giving:

    let  1  2
1.A   A 10  5
1.B   B 11  5
1.C   C 21 10

3) reshape2 Using let and num from above:

library(reshape2)

dcast(let ~ num, data = data.frame(value = unlist(DF)), value.var = "value")

giving:

  let  1  2
1   A 10  5
2   B 11  5
3   C 21 10

Note

The input in reproducible form:

Lines <- "
A1 A2 B1 B2  C1  C2
10  5 11  5  21  10"
DF <- read.table(text = Lines, header = TRUE)

Answer 3

A data.table solution:

library(data.table)
library(magrittr)
melt(df1, measure.vars = names(df1)) %>%
  .[, c("l", "n") := tstrsplit(variable, "")] %>%
  dcast(l ~ n)

   l  1  2
1: A 10  5
2: B 11  5
3: C 21 10