在 R 中创建一个包含嵌套列的表

Question

I have a table in R that is built using this data frame:

Year <- c('2016','2017','2018','2016','2017','2018')
D <- c(0.6,0.57,.56,.41,.45,.85)
C <- c(.8,.9,.65,.8,.9,.65)
var <- c('a','a','a','b','b','b')
dat <- cbind.data.frame(var,Year,D,C)

My goal create a summary table with a header row for each year, then nest(apologies if this not the correct word here) fields D and E under said year: desired output

Answer 1

You can't have a multi-row index in R. This might work:

reshape(dat, idvar = "var", timevar = "Year", direction = "wide")

    var D.2016 C.2016 D.2017 C.2017 D.2018 C.2018
1   a   0.60    0.8   0.57    0.9   0.56   0.65
4   b   0.41    0.8   0.45    0.9   0.85   0.65

Answer 2

An option with pivot_wider

library(dplyr)
library(tidyr)
dat %>% 
   pivot_wider(names_from = Year, values_from = c(D, C))
# A tibble: 2 x 7
#  var   D_2016 D_2017 D_2018 C_2016 C_2017 C_2018
#  <chr>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>
#1 a       0.6   0.570   0.56    0.8    0.9   0.65
#2 b       0.41  0.45    0.85    0.8    0.9   0.65

---

Or using dcast from data.table

library(data.table)
dcast(setDT(dat), var ~ Year, value.var = c("D", "C"))
#   var D_2016 D_2017 D_2018 C_2016 C_2017 C_2018
#1:   a   0.60   0.57   0.56    0.8    0.9   0.65
#2:   b   0.41   0.45   0.85    0.8    0.9   0.65

Answer 3

Using base R, it is not quite true that you can produce the desired output. R dataframes do not allow for MultiIndex unless you use ftables . The closses you could get is using dataframes is provided by the below solutions. If you dont mind using ftables you could do:

tm <- c("D", "C") # the columns to change
a <- reshape(dat, list(tm), idvar = c("var", "Year"), dir = "long", times = tm))
ftable(xtabs(D~., a), row.vars = 1)

    Year 2016      2017      2018     
    time    C    D    C    D    C    D
var                                   
a        0.80 0.60 0.90 0.57 0.65 0.56
b        0.80 0.41 0.90 0.45 0.65 0.85

EDIT: since you are using the KableExtra package, You could do:

nm <- setNames(rep(1:2, c(2, length(unique(dat$Year)))), c(rep(" ", 2), levels(dat$Year)))

library(kableExtra)
dat %>%
  reshape(idvar = "var", timevar = "Year", direction = "wide") %>%
  setNames(gsub("[0-9.]+", "", names(.))) %>%
  kable() %>%
  kable_styling("striped") %>%
  add_header_above(nm)

The result: