重塑但擴展R中的數據

Question

我有以下數據集：

my.data <- read.table(text = '
                  ID  tmc_code  wDay    time_category   TTTR
                  1  121-04711  weekday Afternoon   1.1
                  2  121-04711  weekend Evening     1.3
                  3  121-04711  weekday Morning 1.1
                  4  121-04712  weekend Afternoon   1.101626016
                  5  121-04712  weekday Evening 1.281124498
                  6  121-04712  weekday Morning 1.080645161
                  ', header = TRUE, stringsAsFactors = FALSE, na.strings = 'NA')
my.data

我希望有一個像這樣的寬格式結果：

#result
  #          tmc_code    wDay    TTTR_afternnon TTTR_Evening  TTTR_Morning
  #          121-04711  weekday         1.1         1.3           NA
  #          121-04711  weekend         NA          NA            1.1
  #          121-04712  weekday         NA       1.281124498    1.080645161
  #          121-04712  weekend    1.101626016      NA            NA

我們可以看到帽子不僅應該使用重塑功能，而且實際上這個過程將把6個數據變成9個數據。

下面的重塑功能不適用於這種情況：

w.my.data <- reshape(my.data, idvar = "tmc_code", timevar = "time_category", direction = "wide")

我想知道誰有更好的想法？ 非常感激！

Answer 1

您可以使用reshape2包：

> reshape2::dcast(my.data, tmc_code + wDay ~ paste("TTTR", time_category, sep="_"))

Using TTTR as value column: use value.var to override.
   tmc_code    wDay TTTR_Afternoon TTTR_Evening TTTR_Morning
1 121-04711 weekday       1.100000           NA     1.100000
2 121-04711 weekend             NA     1.300000           NA
3 121-04712 weekday             NA     1.281124     1.080645
4 121-04712 weekend       1.101626           NA           NA

哦，顯然它也適用於reshape ，這也提供了一個有用的警告，關於ID的變化在這里被忽略：

> reshape(my.data, idvar = c("tmc_code", "wDay"), timevar = "time_category", v.names = "TTTR", direction = "wide")

   ID  tmc_code    wDay TTTR.Afternoon TTTR.Evening TTTR.Morning
1:  1 121-04711 weekday       1.100000           NA     1.100000
2:  2 121-04711 weekend             NA     1.300000           NA
3:  4 121-04712 weekend       1.101626           NA           NA
4:  5 121-04712 weekday             NA     1.281124     1.080645
Warning message:
In reshapeWide(data, idvar = idvar, timevar = timevar, varying = varying,  :
  some constant variables (ID) are really varying

Answer 2

類似於tidyr::spread的答案，但使用tidyr::spread ：

library(tidyverse)

my.data %>% 
  select(-ID) %>% # Be sure no important info is lost/misrepresented in dropping ID
  mutate(time_category = paste0("TTTR", "_", time_category)) %>%
  spread(time_category, TTTR)

   tmc_code    wDay TTTR_Afternoon TTTR_Evening TTTR_Morning
1 121-04711 weekday       1.100000           NA     1.100000
2 121-04711 weekend             NA     1.300000           NA
3 121-04712 weekday             NA     1.281124     1.080645
4 121-04712 weekend       1.101626           NA           NA