reshape R 中的数据(宽 -> 长)

Question

我想将df1转换成df2

旧示例数据框 df1

df1 <- structure(list(ID = 1:2,                Group = c(1L, 1L),
                      M1a2hB = c(0.2, 0.3),    M1a3hB = c(0.4, 0.6),
                      M2a2hB = c(0.3, 0.4),    M2a3hB = c(0.6, 0.6),
                      M1r2hB = c(200L, 300L),  M1r3hB = c(400L, 600L),
                      M2r2hB = c(300L, 400L),  M2r3hB = c(600L, 600L)),
                 .Names = c("ID", "Group", "M1a2hB", "M1a3hB", "M2a2hB",
                            "M2a3hB","M1r2hB", "M1r3hB","M2r2hB", "M2r3hB"),
                 class = "data.frame", row.names = c(NA, -2L))

ID Group M1a2hB M1a3hB M2a2hB M2a3hB.... M1r2hB M1r3hB M2r2hB M2r3hB ...
1   1      0.2  0.4    0.3   0.6    ...     200    400   300    600    ...
2   1      0.3  0.6    0.4   0.6    ...     300    600   400    600    ...

这里，df1 有 100 个 ID 和 1100 个列。每个结果 measure 都有两列用于绝对变化，两列用于相对变化。有将近 270 个结果 m 措施。

M1a2hB 是第一次测量从时间 2 到基线的绝对变化，M1a3hB 是时间 3 到基线的绝对变化。同样，M1r2hB 是第一个结果从时间 2 到基线的相对变化，M1r3hB 是结果从时间 3 到基线的相对变化。

新df2:

ID Group time  M1a           M2a        ...  M1r           M2r        ...
1  1     1     0.0           0.0        ...  000           000         ...
1  1     2     0.2           0.3        ...  200           300         ...
1  1     3     0.4           0.6        ...  400           600         ...
2  1     1     0.0           0.0        ...  000           000         ...
2  1     2     0.3           0.4        ...  300           400         ...
2  1     3     0.6           0.6        ...  600           600         ...

有什么建议吗？随时要求任何澄清。谢谢!期待!

附注我尝试运行以前帖子中的一些代码(如果有兴趣请看下面)，但它们看起来不同，因为 df 是三维数据，而 df2 包含额外的时间列

In R, plotting wide form data with ggplot2 or base plot. Is there a way to use ggplot2 without melting wide form data frame?

Reshaping repeated measures data in R wide to long

Answer 1

我们可以使用 sub 从列名中提取模式，split 该向量的序列与 'nm1'，将其用作 measure 在 melt 中将“宽”格式转换为“长”格式。

library(data.table)
nm1 <- sub("\\d+[[:alpha:]]+$", '', names(df1)[-(1:2)])
lst <- split(seq_along(nm1)+2, nm1)
melt(setDT(df1), measure = lst, 
       value.name= names(lst), variable.name= 'time')[order(ID)]
#   ID Group time M1a M1r M2a M2r
#1:  1     1    1 0.2 200 0.3 300
#2:  1     1    2 0.4 400 0.6 600
#3:  2     1    1 0.3 300 0.4 400
#4:  2     1    2 0.6 600 0.6 600

数据

df1 <- structure(list(ID = 1:2, Group = c(1L, 1L),
  M1a2hB = c(0.2, 0.3
), M1a3hB = c(0.4, 0.6), M2a2hB = c(0.3, 0.4),
 M2a3hB = c(0.6, 
0.6), M1r2hB = c(200L, 300L), M1r3hB = c(400L, 600L), 
M2r2hB = c(300L, 
400L), M2r3hB = c(600L, 600L)), .Names = c("ID", "Group", "M1a2hB", 
"M1a3hB", "M2a2hB", "M2a3hB", "M1r2hB", "M1r3hB",
"M2r2hB", "M2r3hB"
), class = "data.frame", row.names = c(NA, -2L))