通过分隔列融化数据

Question

I have a data frame df resulted from dplyr's summarise_each

 V1_mean V2_mean V3_mean V4_mean V5_mean V1_median V2_median V3_median V4_median V5_median V1_my_mode V2_my_mode V3_my_mode V4_my_mode V5_my_mode V1_sum
1       3       4      NA    3.75       5         3         4        NA         4         5          1          2         NA          4          5      4
  V2_sum V3_sum V4_sum V5_sum
1      4      4      4      4

How can I put it in a following format?

  var mean median my_mode sum
1  V1    3      3       1   4
2  V2    4      4       2   4
3  V3   NA     NA      NA   4
4  V4    3      4       4   4
5  V5    5      5       5   4

df

structure(list(V1_mean = 3, V2_mean = 4, V3_mean = NA_real_, 
    V4_mean = 3.75, V5_mean = 5, V1_median = 3, V2_median = 4, 
    V3_median = NA_real_, V4_median = 4, V5_median = 5, V1_my_mode = "1", 
    V2_my_mode = "2", V3_my_mode = NA, V4_my_mode = "4", V5_my_mode = "5", 
    V1_sum = 4L, V2_sum = 4L, V3_sum = 4L, V4_sum = 4L, V5_sum = 4L), class = "data.frame", .Names = c("V1_mean", 
"V2_mean", "V3_mean", "V4_mean", "V5_mean", "V1_median", "V2_median", 
"V3_median", "V4_median", "V5_median", "V1_my_mode", "V2_my_mode", 
"V3_my_mode", "V4_my_mode", "V5_my_mode", "V1_sum", "V2_sum", 
"V3_sum", "V4_sum", "V5_sum"), row.names = c(NA, -1L))

Answer 1

Here's an approach that makes one small modification: Changes my_mode to myMode in the column names. This makes it easy for separate to work after having melt ed the data (using gather ):

library(dplyr)
library(tidyr)
df %>%
  gather(var, val, starts_with("V")) %>%
  separate(var, into = c("V1", "V2")) %>%
  spread(V2, val)
#   V1 mean median myMode sum
# 1 V1    3      3      1   4
# 2 V2    4      4      2   4
# 3 V3 <NA>   <NA>   <NA>   4
# 4 V4 3.75      4      4   4
# 5 V5    5      5      5   4