如何将具有不同列名的具有不同列名的数据帧从宽到长转换

Question

I have a data frame in wide format that I want to transform to long format (melting) so I can process it. The problem is that the "P" columns have different names and the new data frame needs a new "Channel" column so that no information from the header is lost. Please see image below for a pictorial.

Here is the data frame:

df <- read.table(text=
"ID    T    P.1 P.2 P.3
1   24.3    10.2    5.5 2.1
2   23.4    10.4    5.7 2.8
3   22.1    10.5    5.9 3.1
4   19.9    10.2    5.2 2.4
", header=T)

Answer 1

This is a fairly straightforward "wide" to "long" problem. Here are three approaches:

With "reshape2"

library(reshape2)
melt(df, id.vars = c("ID", "T"), variable.name = "Channel", value.name = "P")
#    ID    T Channel    P
# 1   1 24.3     P.1 10.2
# 2   2 23.4     P.1 10.4
# 3   3 22.1     P.1 10.5
# 4   4 19.9     P.1 10.2
# 5   1 24.3     P.2  5.5
# 6   2 23.4     P.2  5.7
# 7   3 22.1     P.2  5.9
# 8   4 19.9     P.2  5.2
# 9   1 24.3     P.3  2.1
# 10  2 23.4     P.3  2.8
# 11  3 22.1     P.3  3.1
# 12  4 19.9     P.3  2.4

With base R's reshape

reshape(df, direction = "long", 
        idvar = c("ID", "T"), 
        timevar = "Channel", 
        varying = 3:ncol(df))
#          ID    T Channel    P
# 1.24.3.1  1 24.3       1 10.2
# 2.23.4.1  2 23.4       1 10.4
# 3.22.1.1  3 22.1       1 10.5
# 4.19.9.1  4 19.9       1 10.2
# 1.24.3.2  1 24.3       2  5.5
# 2.23.4.2  2 23.4       2  5.7
# 3.22.1.2  3 22.1       2  5.9
# 4.19.9.2  4 19.9       2  5.2
# 1.24.3.3  1 24.3       3  2.1
# 2.23.4.3  2 23.4       3  2.8
# 3.22.1.3  3 22.1       3  3.1
# 4.19.9.3  4 19.9       3  2.4

With "tidyr" + "dplyr"

library(dplyr)
library(tidyr)

df %>%
  gather(Channel, P, P.1:P.3) %>%
  mutate(Channel = gsub("P.", "", Channel))
#    ID    T Channel    P
# 1   1 24.3       1 10.2
# 2   2 23.4       1 10.4
# 3   3 22.1       1 10.5
# 4   4 19.9       1 10.2
# 5   1 24.3       2  5.5
# 6   2 23.4       2  5.7
# 7   3 22.1       2  5.9
# 8   4 19.9       2  5.2
# 9   1 24.3       3  2.1
# 10  2 23.4       3  2.8
# 11  3 22.1       3  3.1
# 12  4 19.9       3  2.4

Answer 2

reshape(df,direction="long", varying=list(names(df)[3:5]), v.names="Value",idvar=c("ID","T"))
         ID    T time Value
1.24.3.1  1 24.3    1  10.2
2.23.4.1  2 23.4    1  10.4
3.22.1.1  3 22.1    1  10.5
4.19.9.1  4 19.9    1  10.2
1.24.3.2  1 24.3    2   5.5
2.23.4.2  2 23.4    2   5.7
3.22.1.2  3 22.1    2   5.9
4.19.9.2  4 19.9    2   5.2
1.24.3.3  1 24.3    3   2.1
2.23.4.3  2 23.4    3   2.8
3.22.1.3  3 22.1    3   3.1
4.19.9.3  4 19.9    3   2.4