如何将一个函数返回的多个列突变为R中的另一个数据帧？

Question

我有以下数据帧songs

|artistName   |trackName                 |id                     |duration |
|:------------|:-------------------------|:----------------------|:--------|
|Shaan        |Woh Ladki Hai Kahan       |2tO0QPdCA0jgDBshTuYYkc |306133   |
|Amit Trivedi |Zinda                     |1gEHNfJRSXpDVaEicwRRfe |241962   |
|flora cash   |You're Somebody Else      |0pdKRp2sUthTPe7RLWpPqQ |218883   |
|Iron & Wine  |Each Coming Night         |2xa9PoO42d7VjT0KqG5d3I |207773   |
|Yasser Desai |Tenu Na Bol Pawaan        |5mp7og60TpyNiKi2p9Morw |295352   |
|ERock        |Jurassic Park Meets Metal |4u0RJ8g7QM5exHIPvJ0WW3 |244000   |

我对id专栏很感兴趣。它由{spotifyr}包中的get_track_audio_features()函数用来返回一个音轨音频特性的数据帧。例如，一个id可能返回：

| danceability| energy| key| loudness| mode| speechiness| acousticness| instrumentalness| liveness| valence|  tempo|type           |id                     |uri                                  |track_href                                               |analysis_url                                                     | duration_ms| time_signature|
|------------:|------:|---:|--------:|----:|-----------:|------------:|----------------:|--------:|-------:|------:|:--------------|:----------------------|:------------------------------------|:--------------------------------------------------------|:----------------------------------------------------------------|-----------:|--------------:|
|         0.47|  0.237|   2|  -15.593|    1|      0.0348|        0.837|            0.464|    0.115|   0.358| 81.268|audio_features |2xa9PoO42d7VjT0KqG5d3I |spotify:track:2xa9PoO42d7VjT0KqG5d3I |https://api.spotify.com/v1/tracks/2xa9PoO42d7VjT0KqG5d3I |https://api.spotify.com/v1/audio-analysis/2xa9PoO42d7VjT0KqG5d3I |      207773|              4|

我想做的是添加某些音频功能，例如energy和danceability到songs中，对应于它们的is。生成的数据帧变成：

| id                     | artistName   | trackName            | duration | energy | danceability |
|------------------------|--------------|----------------------|----------|--------|--------------|
| 0pdKRp2sUthTPe7RLWpPqQ | flora cash   | You're Somebody Else | 2349023  | .47    | .13          |
| 1gEHNfJRSXpDVaEicwRRfe | Amit Trivedi | Zinda                | 234009   | .15    | .78          |

我想要使用rowwise()函数来实现这一点，并将所有内容都变成一个单独的列，然后将其分开(因为它们将用逗号分隔，并且可以变成新的列)：

song_data <- songs %>%
    rowwise() %>%
    mutate(song_data_id = toString(get_track_audio_features(id)) %>%
    separate(.,
             song_data_id,
             into =  c("danceablity", "energy"),
             sep = ",")

我认为这是可行的，因为函数返回的前两列是danceability和energy，其余的将被丢弃。但是，它并没有这样做，而是给出了以下错误：

cannot coerce type 'closure' to vector of type 'character'

如何才能以最短的方式将多个列变异到我的数据帧中？

Dput：

structure(list(endTime = structure(c(1588582680, 1588640040, 
1588640040, 1588640280, 1588640280, 1588641240), class = c("POSIXct", 
"POSIXt"), tzone = ""), artistName = c("Shaan", "Amit Trivedi", 
"flora cash", "Amit Trivedi", "flora cash", "flora cash"), trackName = c("Woh Ladki Hai Kahan", 
"Zinda", "You're Somebody Else", "Zinda", "You're Somebody Else", 
"You're Somebody Else"), msPlayed = c(17253L, 0L, 218649L, 1504L, 
218883L, 638496L), date = structure(c(18386, 18387, 18387, 18387, 
18387, 18387), class = "Date"), time = structure(c(52080, 23040, 
23040, 23280, 23280, 24240), class = c("hms", "difftime"), units = "secs")), row.names = c(NA, 
6L), class = "data.frame")

Answer 1

由于该函数返回一个数据帧，因此我们可以使用{purrr}中的map函数将其作为列表存储到列中。map()总是返回一个列表。

features <- song_data %>% 
  nest(data = c(id)) %>%
  mutate(audio_features = map(data, ~ get_track_audio_features(.$id))) %>% 
  unnest(., audio_features) 

然后，音频特征被存储在一个名为audio_features的列中，但这仍然是一个列表。最后一行，unnest()函数将列表展开为单独的列。全都做完了!

Answer 2

如果您将id添加到get_track_audio_features()的结果中，则可以使用join组合表，例如left_join或inner_join。