将多个值列重塑为宽格式

Question

我有以下数据框，想要使用cast函数创建一个"数据透视表"，其中包含两个值的列（value和percent）。

下面是数据框：

expensesByMonth <- structure(list(month = c("2012-02-01", "2012-02-01", "2012-02-01", 
"2012-02-01", "2012-02-01", "2012-02-01", "2012-02-01", "2012-02-01", 
"2012-02-01", "2012-02-01", "2012-02-01", "2012-02-01", "2012-03-01", 
"2012-03-01", "2012-03-01", "2012-03-01", "2012-03-01", "2012-03-01", 
"2012-03-01", "2012-03-01", "2012-03-01", "2012-03-01", "2012-03-01", 
"2012-03-01", "2012-03-01", "2012-03-01", "2012-03-01", "2012-04-01", 
"2012-04-01", "2012-04-01", "2012-04-01", "2012-04-01", "2012-04-01", 
"2012-04-01", "2012-04-01", "2012-04-01", "2012-04-01", "2012-04-01", 
"2012-04-01", "2012-04-01", "2012-04-01", "2012-04-01", "2012-04-01", 
"2012-04-01", "2012-04-01", "2012-05-01", "2012-05-01", "2012-05-01", 
"2012-05-01", "2012-05-01", "2012-05-01", "2012-05-01", "2012-05-01", 
"2012-05-01", "2012-05-01", "2012-05-01", "2012-05-01", "2012-05-01", 
"2012-05-01", "2012-05-01", "2012-05-01", "2012-05-01", "2012-05-01", 
"2012-06-01", "2012-06-01", "2012-06-01", "2012-06-01", "2012-06-01", 
"2012-06-01", "2012-06-01", "2012-06-01", "2012-06-01", "2012-06-01", 
"2012-06-01", "2012-06-01", "2012-06-01", "2012-06-01", "2012-06-01", 
"2012-06-01", "2012-06-01", "2012-06-01", "2012-06-01", "2012-06-01", 
"2012-07-01", "2012-07-01", "2012-07-01", "2012-07-01", "2012-07-01", 
"2012-07-01", "2012-07-01", "2012-07-01", "2012-07-01", "2012-07-01", 
"2012-07-01", "2012-07-01", "2012-07-01"), 
expense_type = c("Adjustment", "Bank Service Charge", "Cable", "Clubbing", "Dining", "Education", 
"Gifts", "Groceries", "Lunch", "Personal Care", "Rent", "Transportation", 
"Adjustment", "Bank Service Charge", "Cable", "Clubbing", "Dining", 
"Gifts", "Groceries", "Lunch", "Medical Expenses", "Miscellaneous", 
"Personal Care", "Phone", "Recreation", "Rent", "Transportation", 
"Adjustment", "Bank Service Charge", "Clothes", "Clubbing", "Computer", 
"Dining", "Gifts", "Groceries", "Lunch", "Maintenance", "Medical Expenses", 
"Miscellaneous", "Personal Care", "Phone", "Recreation", "Rent", 
"Transportation", "Travel", "Bank Service Charge", "Cable", "Clothes", 
"Clubbing", "Computer", "Dining", "Electric", "Gifts", "Groceries", 
"Lunch", "Maintenance", "Medical Expenses", "Miscellaneous", 
"Personal Care", "Phone", "Recreation", "Rent", "Transportation", 
"Adjustment", "Bank Service Charge", "Cable", "Charity", "Clothes", 
"Computer", "Dining", "Education", "Electric", "Gifts", "Groceries", 
"Lunch", "Maintenance", "Medical Expenses", "Miscellaneous", 
"Personal Care", "Phone", "Recreation", "Rent", "Transportation", 
"Computer", "Gifts", "Groceries", "Lunch", "Maintenance", "Medical Expenses", 
"Miscellaneous", "Personal Care", "Phone", "Recreation", "Rent", 
"Repair and Maintenance", "Transportation"), 
value = c(442.37, 200, 21.33, 75, 22.5, 1800, 10, 233.33, 154.75, 30, 545, 32.5, 
2, 200, 36.33, 206.55, 74.5, 89, 372.68, 383.75, 144.19, 508.11, 
30, 38.4, 81.75, 1746.7, 35, 16.37, 200, 806.9, 324.81, 756, 
80.5, 100, 398.37, 326.25, 151, 29.95, 101, 90, 38.45, 61, 743.75, 
129, 228.53, 200, 39.05, 237, 40, 283.83, 141.32, 32.88, 30, 
424.4, 412, 142.75, 86.55, 1051.5, 30, 38.9, 51.5, 749.7, 35, 
10, 200, 16, 32.59, 149.81, 100, 80, 60, 31.91, 55, 397.25, 486.4, 
115.6, 47.08, 1000, 120, 41.11, 256, 761.6, 55, 10.54, 10, 342.11, 
291, 76.5, 66.8, 1008, 30, 41.11, 316, 765, 65, 62), 
percent = c(0.124025030980324, 0.0560729845967511, 0.00598018380724351, 0.0210273692237817, 
0.0063082107671345, 0.50465686137076, 0.00280364922983756, 0.0654175474797997, 
0.0433864718317362, 0.00841094768951267, 0.152798883026147, 0.00911185999697206, 
0.000506462461002391, 0.0506462461002391, 0.00919989060410842, 
0.0523049106600219, 0.018865726672339, 0.0225375795146064, 0.0943742149831854, 
0.0971774847048337, 0.0365134111259673, 0.128669320529962, 0.00759693691503586, 
0.0097240792512459, 0.0207016530934727, 0.442318990316438, 0.00886309306754183, 
0.00357276925628781, 0.0436502047194601, 0.176106750940662, 0.0708901149746392, 
0.164997773839559, 0.0175692073995827, 0.0218251023597301, 0.0869446602704567, 
0.0712043964486193, 0.0329559045631924, 0.00653661815673915, 
0.0220433533833274, 0.0196425921237571, 0.00839175185731621, 
0.0133133124394353, 0.162324198800492, 0.0281543820440518, 0.0498769064226911, 
0.0496724104530621, 0.00969853814096037, 0.0588618063868785, 
0.00993448209061241, 0.070492601294463, 0.0350985252261336, 0.0081661442784834, 
0.00745086156795931, 0.105404854981398, 0.102325165533308, 0.035453682960873, 
0.0214957356235626, 0.261152697956974, 0.00745086156795931, 0.00966128383312057, 
0.0127906456916635, 0.186197030583303, 0.00869267182928586, 0.00249044292527426, 
0.0498088585054852, 0.00398470868043882, 0.00811635349346881, 
0.0373093254635337, 0.0249044292527426, 0.0199235434021941, 0.0149426575516456, 
0.00794700337455016, 0.0136974360890084, 0.09893284520652, 0.12113514388534, 
0.0287895202161704, 0.0117250052921912, 0.249044292527426, 0.0298853151032911, 
0.0102382108658025, 0.0637553388870211, 0.189672133188888, 0.0136974360890084, 
0.00341757293956667, 0.0032424790697976, 0.110928451456846, 0.0943561409311103, 
0.0248049648839517, 0.021659760186248, 0.326841890235599, 0.00972743720939281, 
0.013329831455938, 0.102462338605604, 0.248049648839517, 0.0210761139536844, 
0.0201033702327451)), 
.Names = c("month", "expense_type", "value", "percent"), 
row.names = c(NA, -96L), 
class = "data.frame"
)

这就是我想要创建的内容（当然，使用不同的标题名称，如：[月份]_数值，[月份]_百分比）：

expenses   value     percent value.1   percent.1 value.2   percent.2 value.3   percent.3 value.4   percent.4 value.5   percent.5
1              Adjustment  442.37 0.124025031    2.00 0.000506462   16.37 0.003572769    0.00 0.000000000   10.00 0.002490443    0.00 0.000000000
2     Bank Service Charge  200.00 0.056072985  200.00 0.050646246  200.00 0.043650205  200.00 0.049672410  200.00 0.049808859    0.00 0.000000000
3                   Cable   21.33 0.005980184   36.33 0.009199891    0.00 0.000000000   39.05 0.009698538   16.00 0.003984709    0.00 0.000000000
4                 Charity    0.00 0.000000000    0.00 0.000000000    0.00 0.000000000    0.00 0.000000000   32.59 0.008116353    0.00 0.000000000
5                 Clothes    0.00 0.000000000    0.00 0.000000000  806.90 0.176106751  237.00 0.058861806  149.81 0.037309325    0.00 0.000000000
6                Clubbing   75.00 0.021027369  206.55 0.052304911  324.81 0.070890115   40.00 0.009934482    0.00 0.000000000    0.00 0.000000000
7                Computer    0.00 0.000000000    0.00 0.000000000  756.00 0.164997774  283.83 0.070492601  100.00 0.024904429   10.54 0.003417573
8                  Dining   22.50 0.006308211   74.50 0.018865727   80.50 0.017569207  141.32 0.035098525   80.00 0.019923543    0.00 0.000000000
9               Education 1800.00 0.504656861    0.00 0.000000000    0.00 0.000000000    0.00 0.000000000   60.00 0.014942658    0.00 0.000000000
10               Electric    0.00 0.000000000    0.00 0.000000000    0.00 0.000000000   32.88 0.008166144   31.91 0.007947003    0.00 0.000000000
11                  Gifts   10.00 0.002803649   89.00 0.022537580  100.00 0.021825102   30.00 0.007450862   55.00 0.013697436   10.00 0.003242479
12              Groceries  233.33 0.065417547  372.68 0.094374215  398.37 0.086944660  424.40 0.105404855  397.25 0.098932845  342.11 0.110928451
13                  Lunch  154.75 0.043386472  383.75 0.097177485  326.25 0.071204396  412.00 0.102325166  486.40 0.121135144  291.00 0.094356141
14            Maintenance    0.00 0.000000000    0.00 0.000000000  151.00 0.032955905  142.75 0.035453683  115.60 0.028789520   76.50 0.024804965
15       Medical Expenses    0.00 0.000000000  144.19 0.036513411   29.95 0.006536618   86.55 0.021495736   47.08 0.011725005   66.80 0.021659760
16          Miscellaneous    0.00 0.000000000  508.11 0.128669321  101.00 0.022043353 1051.50 0.261152698 1000.00 0.249044293 1008.00 0.326841890
17          Personal Care   30.00 0.008410948   30.00 0.007596937   90.00 0.019642592   30.00 0.007450862  120.00 0.029885315   30.00 0.009727437
18                  Phone    [phone]   38.40 0.009724079   38.45 0.008391752   38.90 0.009661284   41.11 0.010238211   41.11 0.013329831
19             Recreation    0.00 0.000000000   81.75 0.020701653   61.00 0.013313312   51.50 0.012790646  256.00 0.063755339  316.00 0.102462339
20                   Rent  545.00 0.152798883 1746.70 0.442318990  743.75 0.162324199  749.70 0.186197031  761.60 0.189672133  765.00 0.248049649
21 Repair and Maintenance    0.00 0.000000000    0.00 0.000000000    0.00 0.000000000    0.00 0.000000000    0.00 0.000000000   65.00 0.021076114
22         Transportation   32.50 0.009111860   35.00 0.008863093  129.00 0.028154382   35.00 0.008692672   55.00 0.013697436   62.00 0.020103370
23                 Travel    0.00 0.000000000    0.00 0.000000000  228.53 0.049876906    0.00 0.000000000    0.00 0.000000000    0.00 0.000000000

在将cast应用于单个值列时，我也遇到了以下错误：它没有考虑"value"参数。因此，即使我指定value = "percent"，它仍然显示来自"value"列的值。

cast(expensesByMonth, expense_type ~ month, fun.aggregate = sum, value = "percent")

- Alex Burdusel

---

下面的答案非常有帮助，谢谢。不过我想知道：这个任务能否通过tidyr完成？ - fredtal

Answer 1

data.table 可以对多个 value.var 变量进行转换。这非常直接（而且高效）。

因此：

library(data.table) # v1.9.5+
dcast(setDT(expensesByMonth), expense_type ~ month, value.var = c("value", "percent"))

- Arun

Answer 2

您最好的选择是使用 melt 将数据转换为长格式，然后再使用 dcast：

library(reshape2)

meltExpensesByMonth <- melt(expensesByMonth, id.vars=1:2)
dcast(meltExpensesByMonth, expense_type ~ month + variable, fun.aggregate = sum)

输出的前几行：

             expense_type 2012-02-01_value 2012-02-01_percent 2012-03-01_value 2012-03-01_percent
1              Adjustment           442.37        0.124025031             2.00       0.0005064625
2     Bank Service Charge           200.00        0.056072985           200.00       0.0506462461
3                   Cable            21.33        0.005980184            36.33       0.0091998906
4                 Charity             0.00        0.000000000             0.00       0.0000000000

- Andrie

---

谢谢，这正是我想要的。另外，您有什么想法为什么在这个上面强制转换没有按预期工作吗？cast（expensesByMonth，expense_type〜month，fun.aggregate = sum，value =“percent”） - Alex Burdusel

“cast” 是在（现已废弃的）“reshape”包中的一个函数。它已经被“reshape2”中的“dcast”和“acast”所取代。我不再安装旧版本。 - Andrie

这个问题/答案是纯艺术。谢谢。 - Pablo Casas

如果我每个月都有相同expense_type的多个条目，那么我该如何处理呢？我不想在加入这些值之前进行计算。 - Martin

Answer 3

使用 tidyr 1.0.0 中引入的新函数 pivot_wider()，可以将多个值/度量列从长格式转换为宽格式。

相较于以前的 gather() 和 spread()，此方法的优越性在于不再丢失属性（例如日期仍然是日期，字符串仍然是字符串）。 pivot_wider()（对应函数：pivot_longer()）与spread()类似。不过，它提供了额外的功能，例如多个值列。为此，参数values_from指示从哪些列中获取值可能会有多个列名。

可以使用参数values_fill填充NA。

library("tidyr")
library("magrittr")

pivot_wider(expensesByMonth, 
            id_cols = expense_type,
            names_from = month,
            values_from = c(value, percent))
#> # A tibble: 23 x 13
#>    expense_type `value_2012-02-~ `value_2012-03-~ `value_2012-04-~
#>    <chr>                   <dbl>            <dbl>            <dbl>
#>  1 Adjustment              442.               2               16.4
#>  2 Bank Servic~            200              200              200  
#>  3 Cable                    21.3             36.3             NA  
#>  4 Clubbing                 75              207.             325. 
#>  5 Dining                   22.5             74.5             80.5
#>  6 Education              1800               NA               NA  
#>  7 Gifts                    10               89              100  
#>  8 Groceries               233.             373.             398. 
#>  9 Lunch                   155.             384.             326. 
#> 10 Personal Ca~             30               30               90  
#> # ... with 13 more rows, and 9 more variables: `value_2012-05-01` <dbl>,
#> #   `value_2012-06-01` <dbl>, `value_2012-07-01` <dbl>,
#> #   `percent_2012-02-01` <dbl>, `percent_2012-03-01` <dbl>,
#> #   `percent_2012-04-01` <dbl>, `percent_2012-05-01` <dbl>,
#> #   `percent_2012-06-01` <dbl>, `percent_2012-07-01` <dbl>

或者，也可以使用提供更精细控制的pivot spec来进行重塑（请参见下面的链接）：

# see also ?build_wider_spec
spec <- expensesByMonth %>%
    expand(month, .value = c("percent", "value")) %>%
    dplyr::mutate(.name = paste(.$month, .$.value, sep = "_"))
pivot_wider_spec(expensesByMonth, spec = spec)

^{2019年03月26日使用reprex包 (v0.2.1)创建}

另请参见：https://tidyr.tidyverse.org/dev/articles/pivot.html

- hplieninger

Answer 4

由于这个问题经常被访问，我认为还需要一个完整的基本R答案。基本R中的reshape函数非常灵活，也可以轻松应用于此问题：

expenses <- reshape(expensesByMonth, idvar = 'expense_type', direction = 'wide',
                    timevar = 'month', sep = '_')

带有NA值的单元格可以用以下方法替换为0：

expenses[is.na(expenses)] <- 0

按expense_type排序（以便与所需输出进行比较）：

> expenses[order(expenses$expense_type),]
             expense_type value_2012-02-01 percent_2012-02-01 value_2012-03-01 percent_2012-03-01 value_2012-04-01 percent_2012-04-01 value_2012-05-01 percent_2012-05-01 value_2012-06-01 percent_2012-06-01 value_2012-07-01 percent_2012-07-01
1              Adjustment           442.37        0.124025031             2.00       0.0005064625            16.37        0.003572769             0.00        0.000000000            10.00        0.002490443             0.00        0.000000000
2     Bank Service Charge           200.00        0.056072985           200.00       0.0506462461           200.00        0.043650205           200.00        0.049672410           200.00        0.049808859             0.00        0.000000000
3                   Cable            21.33        0.005980184            36.33       0.0091998906             0.00        0.000000000            39.05        0.009698538            16.00        0.003984709             0.00        0.000000000
67                Charity             0.00        0.000000000             0.00       0.0000000000             0.00        0.000000000             0.00        0.000000000            32.59        0.008116353             0.00        0.000000000
30                Clothes             0.00        0.000000000             0.00       0.0000000000           806.90        0.176106751           237.00        0.058861806           149.81        0.037309325             0.00        0.000000000
4                Clubbing            75.00        0.021027369           206.55       0.0523049107           324.81        0.070890115            40.00        0.009934482             0.00        0.000000000             0.00        0.000000000
32               Computer             0.00        0.000000000             0.00       0.0000000000           756.00        0.164997774           283.83        0.070492601           100.00        0.024904429            10.54        0.003417573
5                  Dining            22.50        0.006308211            74.50       0.0188657267            80.50        0.017569207           141.32        0.035098525            80.00        0.019923543             0.00        0.000000000
6               Education          1800.00        0.504656861             0.00       0.0000000000             0.00        0.000000000             0.00        0.000000000            60.00        0.014942658             0.00        0.000000000
52               Electric             0.00        0.000000000             0.00       0.0000000000             0.00        0.000000000            32.88        0.008166144            31.91        0.007947003             0.00        0.000000000
7                   Gifts            10.00        0.002803649            89.00       0.0225375795           100.00        0.021825102            30.00        0.007450862            55.00        0.013697436            10.00        0.003242479
8               Groceries           233.33        0.065417547           372.68       0.0943742150           398.37        0.086944660           424.40        0.105404855           397.25        0.098932845           342.11        0.110928451
9                   Lunch           154.75        0.043386472           383.75       0.0971774847           326.25        0.071204396           412.00        0.102325166           486.40        0.121135144           291.00        0.094356141
37            Maintenance             0.00        0.000000000             0.00       0.0000000000           151.00        0.032955905           142.75        0.035453683           115.60        0.028789520            76.50        0.024804965
21       Medical Expenses             0.00        0.000000000           144.19       0.0365134111            29.95        0.006536618            86.55        0.021495736            47.08        0.011725005            66.80        0.021659760
22          Miscellaneous             0.00        0.000000000           508.11       0.1286693205           101.00        0.022043353          1051.50        0.261152698          1000.00        0.249044293          1008.00        0.326841890
10          Personal Care            30.00        0.008410948            30.00       0.0075969369            90.00        0.019642592            30.00        0.007450862           120.00        0.029885315            30.00        0.009727437
24                  Phone             0.00        0.000000000            38.40       0.0097240793            38.45        0.008391752            38.90        0.009661284            41.11        0.010238211            41.11        0.013329831
25             Recreation             0.00        0.000000000            81.75       0.0207016531            61.00        0.013313312            51.50        0.012790646           256.00        0.063755339           316.00        0.102462339
11                   Rent           545.00        0.152798883          1746.70       0.4423189903           743.75        0.162324199           749.70        0.186197031           761.60        0.189672133           765.00        0.248049649
95 Repair and Maintenance             0.00        0.000000000             0.00       0.0000000000             0.00        0.000000000             0.00        0.000000000             0.00        0.000000000            65.00        0.021076114
12         Transportation            32.50        0.009111860            35.00       0.0088630931           129.00        0.028154382            35.00        0.008692672            55.00        0.013697436            62.00        0.020103370
45                 Travel             0.00        0.000000000             0.00       0.0000000000           228.53        0.049876906             0.00        0.000000000             0.00        0.000000000             0.00        0.000000000

你还可以使用 tidyverse 实现这一点：

library(dplyr)
library(tidyr)

expensesByMonth %>% 
  gather(k, v, 3:4) %>% 
  unite(km, k, month) %>% 
  spread(km, v, fill = 0)

- Jaap

---

我同意，这个答案是必要的。需要注意的是，如果原始数据框中有其他常量变量，则需要使用包含要扩展的数据的所有列的名称的v.names参数进行指定。当然，在这种情况下，这是不必要的，因为它只是所有未指定的列。 - John

Answer 5

我建议使用tables包中的tabulate函数来完成这个任务。该函数需要使用因子，但是对于您拥有的数据类型来说，这本身就是一个好主意。

library(tables)
expensesByMonth$month= as.factor(expensesByMonth$month)
expensesByMonth$expense_type= as.factor(expensesByMonth$expense_type)
tabular(expense_type~(month)*(value+percent)*(sum),data=expensesByMonth)
# Optional formatting
tabular(expense_type~month*
   ((Format(digits=1))*value+(Format(digits=3))*percent)*sum,
   data=expensesByMonth)

部分输出：

                       value      percent  value      percent  value      percent 
expense_type           sum        sum      sum        sum      sum        sum     
Adjustment              442       0.124025    2       0.000506   16       0.003573
Bank Service Charge     200       0.056073  200       0.050646  200       0.043650
Cable                    21       0.005980   36       0.009200    0       0.000000

- Dieter Menne

---

有些()是多余的，但我有时会保留它们以防我想添加更多术语。 - Dieter Menne

我以前从未使用过这个类。如果可以的话，你能否提供一个从中提取数据的示例呢？谢谢！ - Alex Burdusel

这不是一个一行代码的事情。该软件包的作者Duncan Murdoch编写了一个很好的说明文档，附带了tables软件包。请参阅软件包文档/指南和说明概述。 - Dieter Menne

没有简单的方法可以撤销这个过程，以防万一需要（假设）？感谢您的回答。 - Alex Burdusel

不，这是reshape（base）和reshape包中的函数的领域。tabular仅用于显示。 - Dieter Menne