r - 使用 map2_dfr 将数据 bind_cols 在一起

Question

我有两个名为 5 和 6 的列表，如下所示:

df1:

$`5`
$`5`$`2016-01-01`
            [,1]
 [1,] -0.8357399
 [2,]  0.7793535
 [3,]  1.2425095
 [4,]  1.0564501
 [5,]  0.7351215
 [6,] -0.1976808
 [7,]  0.1692951
 [8,] -0.4280740
 [9,] -0.5507262
[10,]  1.0437385

$`5`$`2016-01-02`
             [,1]
 [1,]  0.68302631
 [2,]  1.16508889
 [3,]  1.04583836
 [4,]  0.71979517
 [5,] -0.14236742
 [6,]  0.07792941
 [7,] -0.41616267
 [8,] -0.49082834
 [9,]  1.09706485
[10,]  1.06911635

这是一些预测。我还有另一个列表，如下所示:

df2:

$`5`[[3]]
         date      wind     temp
1  2010-01-03  1.387097 42.09677
2  2010-01-04  3.225806 45.16129
3  2010-01-05  4.192308 47.34615
4  2010-01-06  8.083333 39.83333
5  2010-01-07  8.774194 33.29032
6  2010-01-08  9.709677 32.25806
7  2010-01-09 12.419355 31.16129
8  2010-01-10  1.290323 27.54839
9  2010-01-11  2.258065 31.06452
10 2010-01-12  4.967742 33.83871

$`5`[[4]]
         date      wind     temp
1  2010-01-04  3.225806 45.16129
2  2010-01-05  4.192308 47.34615
3  2010-01-06  8.083333 39.83333
4  2010-01-07  8.774194 33.29032
5  2010-01-08  9.709677 32.25806
6  2010-01-09 12.419355 31.16129
7  2010-01-10  1.290323 27.54839
8  2010-01-11  2.258065 31.06452
9  2010-01-12  4.967742 33.83871
10 2010-01-13  4.129032 40.70968

其中包含一些原始数据。我想将 df1 与 df2 结合起来，这样它看起来像这样:

$`5`[[3]]
         date      wind     temp   prediction
1  2010-01-03  1.387097 42.09677  -0.8357399
2  2010-01-04  3.225806 45.16129   0.7793535
3  2010-01-05  4.192308 47.34615   1.2425095 
4  2010-01-06  8.083333 39.83333   1.0564501
5  2010-01-07  8.774194 33.29032   0.7351215
6  2010-01-08  9.709677 32.25806  -0.1976808
7  2010-01-09 12.419355 31.16129   0.1692951
8  2010-01-10  1.290323 27.54839  -0.4280740
9  2010-01-11  2.258065 31.06452  -0.5507262
10 2010-01-12  4.967742 33.83871   1.0437385

和

$`5`[[4]]
         date      wind     temp   prediction
1  2010-01-04  3.225806 45.16129   0.68302631
2  2010-01-05  4.192308 47.34615   1.16508889
3  2010-01-06  8.083333 39.83333   1.04583836
4  2010-01-07  8.774194 33.29032   0.71979517
5  2010-01-08  9.709677 32.25806  -0.14236742
6  2010-01-09 12.419355 31.16129   0.07792941
7  2010-01-10  1.290323 27.54839  -0.41616267
8  2010-01-11  2.258065 31.06452  -0.49082834
9  2010-01-12  4.967742 33.83871   1.09706485
10 2010-01-13  4.129032 40.70968   1.09706485

最后我想将所有这些列表以长格式绑定(bind)在一起，例如:

       date      wind     temp   prediction  ID
 2010-01-03  1.387097 42.09677  -0.8357399    5
 2010-01-04  3.225806 45.16129   0.7793535    5
 2010-01-05  4.192308 47.34615   1.2425095    5
 2010-01-06  8.083333 39.83333   1.0564501    5
 2010-01-07  8.774194 33.29032   0.7351215    5
 2010-01-08  9.709677 32.25806  -0.1976808    5
 2010-01-09 12.419355 31.16129   0.1692951    5
 2010-01-10  1.290323 27.54839  -0.4280740    5
 2010-01-11  2.258065 31.06452  -0.5507262    5
 2010-01-12  4.967742 33.83871   1.0437385    5
 2010-01-04  3.225806 45.16129   0.68302631   6
 2010-01-05  4.192308 47.34615   1.16508889   6
 2010-01-06  8.083333 39.83333   1.04583836   6
 2010-01-07  8.774194 33.29032   0.71979517   6
 2010-01-08  9.709677 32.25806  -0.14236742   6
 2010-01-09 12.419355 31.16129   0.07792941   6
 2010-01-10  1.290323 27.54839  -0.41616267   6
 2010-01-11  2.258065 31.06452  -0.49082834   6
 2010-01-12  4.967742 33.83871   1.09706485   6
 2010-01-13  4.129032 40.70968   1.09706485   6

这是一个大型的单一数据框，ID 列是列表名称。

我试过以下代码:

library(purrr)
map2_dfr(df1, df2, ~map2_dfr(.x, map(.y, "date"), ~cbind(.x, date = .y))) 

这不符合我的要求。

# A tibble: 40 x 4
   `2016-01-01` `2016-01-02` `2016-01-03` `2016-01-04`
          <dbl>        <dbl>        <dbl>        <dbl>
 1       -0.836       0.683        1.12          0.935
 2        0.779       1.17         1.02          0.754
 3        1.24        1.05         0.776        -0.146
 4        1.06        0.720       -0.188         0.160
 5        0.735      -0.142        0.0884       -0.465
 6       -0.198       0.0779      -0.283        -0.543
 7        0.169      -0.416       -0.453         1.10 
 8       -0.428      -0.491        1.06          1.03 
 9       -0.551       1.10         1.03          0.976
10        1.04        1.07         1.06          1.29 
# … with 30 more rows

数据:

df1 <- list(`5` = list(`2016-01-01` = structure(c(-0.835739850997925, 
0.779353499412537, 1.24250948429108, 1.0564501285553, 0.735121548175812, 
-0.197680771350861, 0.169295132160187, -0.42807400226593, -0.550726175308228, 
1.04373848438263), .Dim = c(10L, 1L)), `2016-01-02` = structure(c(0.683026313781738, 
1.16508889198303, 1.04583835601807, 0.719795167446136, -0.142367422580719, 
0.0779294073581696, -0.416162669658661, -0.490828335285187, 1.09706485271454, 
1.06911635398865), .Dim = c(10L, 1L)), `2016-01-03` = structure(c(1.12009644508362, 
1.01791954040527, 0.775721669197083, -0.187799870967865, 0.0883594155311584, 
-0.283172011375427, -0.452982068061829, 1.06388020515442, 1.02800369262695, 
1.06038355827332), .Dim = c(10L, 1L)), `2016-01-04` = structure(c(0.935240745544434, 
0.753606081008911, -0.145884394645691, 0.160260230302811, -0.464599192142487, 
-0.543198347091675, 1.10469722747803, 1.03225469589233, 0.976013004779816, 
1.28949522972107), .Dim = c(10L, 1L))), `6` = list(`2016-01-01` = structure(c(-1.05108523368835, 
0.831294775009155, 1.0468602180481, 1.4151725769043, 0.89024829864502, 
-0.23750251531601, 0.0968895554542542, -0.447583615779877, -0.885086059570312, 
1.16484880447388), .Dim = c(10L, 1L)), `2016-01-02` = structure(c(0.898594379425049, 
1.03132367134094, 1.43809175491333, 1.05361354351044, -0.204488694667816, 
0.0889829993247986, -0.42036372423172, -0.906104445457458, 1.16317582130432, 
1.13032007217407), .Dim = c(10L, 1L)), `2016-01-03` = structure(c(0.99635636806488, 
1.5137802362442, 0.92145836353302, -0.218990564346313, 0.147298634052277, 
-0.466208696365356, -0.896591305732727, 1.18873286247253, 1.10375666618347, 
1.59670341014862), .Dim = c(10L, 1L)), `2016-01-04` = structure(c(1.45108199119568, 
0.860665202140808, -0.299971401691437, 0.0754360556602478, -0.460747301578522, 
-0.947231769561768, 1.06433939933777, 1.13789772987366, 1.70162570476532, 
2.04307699203491), .Dim = c(10L, 1L))))


df2 <- list(`5` = list(structure(list(date = structure(c(14610, 14611, 
14612, 14613, 14614, 14615, 14616, 14617, 14618, 14619), class = "Date"), 
    wind = c(19.72, 3.19354838709677, 1.38709677419355, 3.2258064516129, 
    4.19230769230769, 8.08333333333333, 8.7741935483871, 9.70967741935484, 
    12.4193548387097, 1.29032258064516), temp = c(41.6, 41.9677419354839, 
    42.0967741935484, 45.1612903225806, 47.3461538461538, 39.8333333333333, 
    33.2903225806452, 32.258064516129, 31.1612903225806, 27.5483870967742
    )), row.names = c(NA, 10L), class = "data.frame"), structure(list(
    date = structure(c(14611, 14612, 14613, 14614, 14615, 14616, 
    14617, 14618, 14619, 14620), class = "Date"), wind = c(3.19354838709677, 
    1.38709677419355, 3.2258064516129, 4.19230769230769, 8.08333333333333, 
    8.7741935483871, 9.70967741935484, 12.4193548387097, 1.29032258064516, 
    2.25806451612903), temp = c(41.9677419354839, 42.0967741935484, 
    45.1612903225806, 47.3461538461538, 39.8333333333333, 33.2903225806452, 
    32.258064516129, 31.1612903225806, 27.5483870967742, 31.0645161290323
    )), row.names = c(NA, 10L), class = "data.frame"), structure(list(
    date = structure(c(14612, 14613, 14614, 14615, 14616, 14617, 
    14618, 14619, 14620, 14621), class = "Date"), wind = c(1.38709677419355, 
    3.2258064516129, 4.19230769230769, 8.08333333333333, 8.7741935483871, 
    9.70967741935484, 12.4193548387097, 1.29032258064516, 2.25806451612903, 
    4.96774193548387), temp = c(42.0967741935484, 45.1612903225806, 
    47.3461538461538, 39.8333333333333, 33.2903225806452, 32.258064516129, 
    31.1612903225806, 27.5483870967742, 31.0645161290323, 33.8387096774194
    )), row.names = c(NA, 10L), class = "data.frame"), structure(list(
    date = structure(c(14613, 14614, 14615, 14616, 14617, 14618, 
    14619, 14620, 14621, 14622), class = "Date"), wind = c(3.2258064516129, 
    4.19230769230769, 8.08333333333333, 8.7741935483871, 9.70967741935484, 
    12.4193548387097, 1.29032258064516, 2.25806451612903, 4.96774193548387, 
    4.12903225806452), temp = c(45.1612903225806, 47.3461538461538, 
    39.8333333333333, 33.2903225806452, 32.258064516129, 31.1612903225806, 
    27.5483870967742, 31.0645161290323, 33.8387096774194, 40.7096774193548
    )), row.names = c(NA, 10L), class = "data.frame")), `6` = list(
    structure(list(date = structure(c(14610, 14611, 14612, 14613, 
    14614, 14615, 14616, 14617, 14618, 14619), class = "Date"), 
        wind = c(19.72, 3.19354838709677, 1.38709677419355, 3.2258064516129, 
        4.19230769230769, 8.08333333333333, 8.7741935483871, 
        9.70967741935484, 12.4193548387097, 1.29032258064516), 
        temp = c(41.6, 41.9677419354839, 42.0967741935484, 45.1612903225806, 
        47.3461538461538, 39.8333333333333, 33.2903225806452, 
        32.258064516129, 31.1612903225806, 27.5483870967742)), row.names = c(NA, 
    10L), class = "data.frame"), structure(list(date = structure(c(14611, 
    14612, 14613, 14614, 14615, 14616, 14617, 14618, 14619, 14620
    ), class = "Date"), wind = c(3.19354838709677, 1.38709677419355, 
    3.2258064516129, 4.19230769230769, 8.08333333333333, 8.7741935483871, 
    9.70967741935484, 12.4193548387097, 1.29032258064516, 2.25806451612903
    ), temp = c(41.9677419354839, 42.0967741935484, 45.1612903225806, 
    47.3461538461538, 39.8333333333333, 33.2903225806452, 32.258064516129, 
    31.1612903225806, 27.5483870967742, 31.0645161290323)), row.names = c(NA, 
    10L), class = "data.frame"), structure(list(date = structure(c(14612, 
    14613, 14614, 14615, 14616, 14617, 14618, 14619, 14620, 14621
    ), class = "Date"), wind = c(1.38709677419355, 3.2258064516129, 
    4.19230769230769, 8.08333333333333, 8.7741935483871, 9.70967741935484, 
    12.4193548387097, 1.29032258064516, 2.25806451612903, 4.96774193548387
    ), temp = c(42.0967741935484, 45.1612903225806, 47.3461538461538, 
    39.8333333333333, 33.2903225806452, 32.258064516129, 31.1612903225806, 
    27.5483870967742, 31.0645161290323, 33.8387096774194)), row.names = c(NA, 
    10L), class = "data.frame"), structure(list(date = structure(c(14613, 
    14614, 14615, 14616, 14617, 14618, 14619, 14620, 14621, 14622
    ), class = "Date"), wind = c(3.2258064516129, 4.19230769230769, 
    8.08333333333333, 8.7741935483871, 9.70967741935484, 12.4193548387097, 
    1.29032258064516, 2.25806451612903, 4.96774193548387, 4.12903225806452
    ), temp = c(45.1612903225806, 47.3461538461538, 39.8333333333333, 
    33.2903225806452, 32.258064516129, 31.1612903225806, 27.5483870967742, 
    31.0645161290323, 33.8387096774194, 40.7096774193548)), row.names = c(NA, 
    10L), class = "data.frame")))

编辑:

我喜欢@tmfmnk 和@akrun 的两个解决方案。

(一些推理 - 主要是为了我稍后再回到这个问题)

在我的完整示例中，我收到错误 Error: Argument 2 must be length 2995920, not 2998110。当我申请时:

map2_dfr(map(df1, ~ bind_rows(.) %>%
              pivot_longer(everything(), values_to = "prediction") %>%
              select(-name)),
         map(df2, bind_rows),
         bind_cols, .id = "ID")

我将错误缩小到以下事实:

map(df1, ~bind_rows(.) %>% 
      pivot_longer(everything(), values_to = "prediction") %>% 
      select(-name))

创建 2 个包含 2995920 观察值的列表。我也分开跑；

map(processed_analysis, ~bind_rows(.))

这会创建 2 个包含 2998110 的列表。这两个数字的区别是 2190(据我所知，这只对我有意义，但它很重要，因为 2190 是列表列表之一的长度或观察值的数量列表列表之一)。

2998110/2190 = 1369 其中 1369 是我在每个列表(5 和 6) 在这个例子中。

@akrun 的解决方案:

imap_dfr(df2, ~ bind_rows(.x) %>%
        mutate(ID = .y)) %>%
    mutate(prediction = unlist(df1)) %>%
    as_tibble

给我一​​个由 5996220 观察值组成的数据框。这与 2998110 + 2998110 或每个列表 5 或 6 包含的观察数相同，当 1369 list of lists 已经rbind。

2998110的@akrun方案和2995920的@tmfmnk方案的区别在于2190，正好是行数的长度5 和 6 的 1369 列表之一。

2998110/1369 = 2190 是正确的，而 2995920/1369 = 2188..44 与数据不一致。

我不明白为什么这两种解决方案似乎对此处提供的数据完美运行，但对我拥有的完整数据却略有不同。

Answer 1

一个选项可能是:

map2_dfr(map(df1, ~ unlist(.) %>%
              enframe(value = "prediction") %>%
              select(-name)),
         map(df2, bind_rows),
         bind_cols, .id = "ID")

   ID    prediction date        wind  temp
   <chr>      <dbl> <date>     <dbl> <dbl>
 1 5         -0.836 2010-01-01 19.7   41.6
 2 5          0.779 2010-01-02  3.19  42.0
 3 5          1.24  2010-01-03  1.39  42.1
 4 5          1.06  2010-01-04  3.23  45.2
 5 5          0.735 2010-01-05  4.19  47.3
 6 5         -0.198 2010-01-06  8.08  39.8
 7 5          0.169 2010-01-07  8.77  33.3
 8 5         -0.428 2010-01-08  9.71  32.3
 9 5         -0.551 2010-01-09 12.4   31.2
10 5          1.04  2010-01-10  1.29  27.5
# … with 70 more rows