使用月份（列）和期间（行）制作数据框

Question

我有一个包含多列月份和总数（总共 13 列）和两个时期（半年，标记为 1 和 7）的数据框。

我正在尝试广泛传播它，并将其呈现为 January1 和 January7（所有 12 个月 + Total 和 Total1）。然后我会计算期间之间的差异。

请指教怎么做。

我尝试了 spread()，老 Hadley 的函数，但是这几个月让键和值的一切变得复杂。也因 pivot_wide().

而失败

显然尝试了多个问题，例如 this。

我的数据样本在这里。

structure(list(Year = c(2019, 2019, 2019, 2019, 2019), Period = c(1, 
1, 7, 1, 7), KPKV = c(99999, 110000, 111000, 111010, 111010), 
KEKV = c(4, 1, 1, 2, 2), Name = c("A", "B", "B", "B", "B"
), January = c(70198346.4, 125181.4, 125181.4, 64008.4, 34374.1
), February = c(71052496.2, 127697.1, 127697.1, 66007.3, 
34719.1), March = c(96884031.5, 142375.3, 142375.3, 75510.2, 
38082.1), April = c(74389605.4, 139627.8, 139627.8, 75891.9, 
37262.5), May = c(101876908, 144649.4, 144649.4, 79889.6, 
41150), June = c(86362730.8, 178706.8, 178706.8, 96616, 49727.9
), July = c(74326532.8, 178708.4, 178708.4, 96616, 55955.7
), August = c(80052666.3, 186225.8, 186225.8, 102606.5, 30816.8
), September = c(90236044.8, 182131, 182131, 102885.7, 49123.1
), October = c(79077964, 175287.8, 175287.8, 101166.1, 49942.8
), November = c(92509081.2, 185182.1, 185182.1, 109051.8, 
37609.2), December = c(88801141.2, 198270.2, 198270.2, 119648, 
37609.2), Total = c(1005767549, 1964043.1, 1964043.1, 1089897.5, 
496372.5)), class = c("spec_tbl_df", "tbl_df", "tbl", "data.frame"
), row.names = c(NA, -5L), spec = structure(list(cols = list(
Year = structure(list(), class = c("collector_double", "collector"
)), Period = structure(list(), class = c("collector_double", 
"collector")), KPKV = structure(list(), class = c("collector_double", 
"collector")), KEKV = structure(list(), class = c("collector_double", 
"collector")), Name = structure(list(), class = c("collector_character", 
"collector")), January = structure(list(), class = c("collector_double", 
"collector")), February = structure(list(), class = c("collector_double", 
"collector")), March = structure(list(), class = c("collector_double", 
"collector")), April = structure(list(), class = c("collector_double", 
"collector")), May = structure(list(), class = c("collector_double", 
"collector")), June = structure(list(), class = c("collector_double", 
"collector")), July = structure(list(), class = c("collector_double", 
"collector")), August = structure(list(), class = c("collector_double", 
"collector")), September = structure(list(), class = c("collector_double", 
"collector")), October = structure(list(), class = c("collector_double", 
"collector")), November = structure(list(), class = c("collector_double", 
"collector")), December = structure(list(), class = c("collector_double", 
"collector")), Total = structure(list(), class = c("collector_double", 
"collector"))), default = structure(list(), class = c("collector_guess", 
"collector")), skip = 1), class = "col_spec"))

更新：

使用第一种解决方案后，数据进行了转换，但并非一切正常。一些列丢失了。我相信这是因为 KPKV 列是唯一的，但是 KEKV 列在同一个 KPKV 下可以有多个值。

我的预期输出

structure(list(Year = 2019, KPKV = 99999, KEKV = 4, Name = "Random name", 
April1 = 74389605.4, April7 = NA_real_, August1 = 80052666.3, 
August7 = NA_real_, December1 = 88801141.2, December7 = NA_real_, 
February1 = 71052496.2, February7 = NA_real_, January1 = 70198346.4, 
January7 = NA_real_, July1 = 74326532.8, July7 = NA_real_, 
June1 = 86362730.8, June7 = NA_real_, March1 = 96884031.5, 
March7 = NA_real_, May1 = 101876908, May7 = NA_real_, November1 = 
92509081.2, 
November7 = NA_real_, October1 = 79077964, October7 = NA_real_, 
September1 = 90236044.8, September7 = NA_real_, Total1 = 1005767548.6, 
Total7 = NA_real_), row.names = 1L, class = "data.frame")

Answer 1

使用基础 R reshape 函数：

 reshape(data.frame(df),idvar = "Name",timevar = "Period",dir="wide",sep="")

  Name Year1  KPKV1 KEKV1   January1  February1     March1     April1        May1      June1      July1    August1 September1
1    A  2019 110000     1 70198346.4 71052496.2 96884031.5 74389605.4 101876908.0 86362730.8 74326532.8 80052666.3   90236045
2    B  2019 110000     1   125181.4   127697.1   142375.3   139627.8    144649.4   178706.8   178708.4   186225.8     182131
    October1  November1  December1     Total1 Year7  KPKV7 KEKV7 January7 February7   March7   April7     May7    June7
1 79077964.0 92509081.2 88801141.2 1005767549  2019 111000     1 125181.4  127697.1 142375.3 139627.8 144649.4 178706.8
2   175287.8   185182.1   198270.2    1964043  2019 111010     1  64008.4   66007.3  75510.2  75891.9  79889.6  96616.0
     July7  August7 September7 October7 November7 December7  Total7
1 178708.4 186225.8   182131.0 175287.8  185182.1  198270.2 1964043
2  96616.0 102606.5   102885.7 101166.1  109051.8  119648.0 1089898

Answer 2

一个选项是 pivot_wider 来自 tidyr

的开发版本

library(tidyr) #‘0.8.3.9000’
library(dplyr)
df1 %>% 
    pivot_wider(id_cols =  Name, names_from = Period, 
       values_from = c(January:December), names_sep = "")
# A tibble: 2 x 25
#  Name  January1 January7 February1 February7 March1 March7 April1 April7   May1   May7  June1  June7  July1  July7 August1 August7 September1 September7
#  <chr>    <dbl>    <dbl>     <dbl>     <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>   <dbl>   <dbl>      <dbl>      <dbl>
#1 A       7.02e7  125181. 71052496.   127697. 9.69e7 1.42e5 7.44e7 1.40e5 1.02e8 #1.45e5 8.64e7 1.79e5 7.43e7 1.79e5  8.01e7 186226.  90236045.    182131 
#2 B       1.25e5   64008.   127697.    66007. 1.42e5 7.55e4 1.40e5 7.59e4 1.45e5 7.99e4 1.79e5 9.66e4 1.79e5 9.66e4  1.86e5 102606.    182131     102886.
# … with 6 more variables: October1 <dbl>, October7 <dbl>, November1 <dbl>, November7 <dbl>, December1 <dbl>, December7 <dbl>

或与以前的 tidyr 版本

library(dplyr)
library(tidyr)
df1 %>% 
    gather(key, val, January:December) %>%
    unite(key, key, Period, sep="") %>%
    spread(key, val)

或data.table选项

library(data.table)
dcast(setDT(df1), Name + KPKV + Year ~ Period, value.var = month.name, sep="")

使用月份（列）和期间（行）制作数据框

Make dataframe wide with months (columns) and periods (rows)

pivot

r

spread

tidyverse