将列表展平为 R 数据框答案

【问题标题】：Flatten list to R dataframe将列表展平为 R 数据框
【发布时间】：2018-11-09 11:46:44
【问题描述】：

我有一个嵌套列表，我想使用 R 将其转换为数据框，类似于这个问题 flatten a data frame

这是我的列表结构

> str(rf_curves$GBP)
List of 27
 $ NA                  :'data.frame':   0 obs. of  2 variables:
  ..$ date   :Class 'Date'  int(0) 
  ..$ px_last: num(0) 
 $ BP0012M       Index :'data.frame':   5 obs. of  2 variables:
  ..$ date   : Date[1:5], format: "2018-05-21" "2018-05-22" ...
  ..$ px_last: num [1:5] 0.929 0.931 0.918 0.918 0.901
 $ BP0003M       Index :'data.frame':   5 obs. of  2 variables:
  ..$ date   : Date[1:5], format: "2018-05-21" "2018-05-22" ...
  ..$ px_last: num [1:5] 0.623 0.623 0.619 0.614 0.611
 $ BP0006M       Index :'data.frame':   5 obs. of  2 variables:
  ..$ date   : Date[1:5], format: "2018-05-21" "2018-05-22" ...
  ..$ px_last: num [1:5] 0.746 0.743 0.734 0.733 0.723
 $ NA                  :'data.frame':   0 obs. of  2 variables:
  ..$ date   :Class 'Date'  int(0) 
  ..$ px_last: num(0)

我想要一个带有

的数据框

日期在行上
列上的 ticker（BP0012M 和 BP0003M 是 tickers 的示例）
填充有px_last的单元格。

因此数据框的样本将是：

date           NA   BP0012M BP0003M BOP0006M
2018-05-21          0.929   0.623   0.746
2018-05-22          0.931   0.623   0.743
2018-05-23          0.918   0.619   0.743
2018-05-24          0.918   0.614   0.733
2018-05-25          0.901   0.611   0.723

最终目标是在给定特定开始日期的情况下找到一种方便的方法来获得无风险曲线。例如 BP0012M 是 12 个月的 GBP libor。我目前使用库（Rblpapi）从彭博社加载数据。如果我可以从另一个提供商那里获得相同的数据，例如Quandl 没关系。如果我能实现这个目标，而无需将列表扁平化为数据框，我也可以使用该解决方案。

编辑：请求的输出粘贴在下面

> dput(rf_curves$GBP)
structure(list(`NA` = structure(list(date = structure(integer(0), class = "Date"), 
    px_last = numeric(0)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = integer(0)), `BP0012M       Index` = structure(list(
    date = structure(17672:17676, class = "Date"), px_last = c(0.92894, 
    0.93081, 0.91831, 0.9182, 0.90056)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 5L)), `BP0003M       Index` = structure(list(
    date = structure(17672:17676, class = "Date"), px_last = c(0.62281, 
    0.6225, 0.619, 0.61406, 0.61067)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 5L)), `BP0006M       Index` = structure(list(
    date = structure(17672:17676, class = "Date"), px_last = c(0.7463, 
    0.74323, 0.73411, 0.73321, 0.72312)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 5L)), `NA` = structure(list(date = structure(integer(0), class = "Date"), 
    px_last = numeric(0)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = integer(0)), `NA` = structure(list(date = structure(integer(0), class = "Date"), 
    px_last = numeric(0)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = integer(0)), `NA` = structure(list(date = structure(integer(0), class = "Date"), 
    px_last = numeric(0)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = integer(0)), `NA` = structure(list(date = structure(integer(0), class = "Date"), 
    px_last = numeric(0)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = integer(0)), `NA` = structure(list(date = structure(integer(0), class = "Date"), 
    px_last = numeric(0)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = integer(0)), `BPSW30   CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.758, 1.768, 
    1.715, 1.696, 1.628, 1.531, 1.56725)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `NA` = structure(list(date = structure(integer(0), class = "Date"), 
    px_last = numeric(0)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = integer(0)), `NA` = structure(list(date = structure(integer(0), class = "Date"), 
    px_last = numeric(0)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = integer(0)), `BPSW8    CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.5675, 1.5955, 
    1.5375, 1.5175, 1.4475, 1.342, 1.37775)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW1F   CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(0.942, 0.9575, 
    0.9225, 0.9188, 0.8864, 0.84505, 0.863)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW9    CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.6115, 1.6395, 
    1.5795, 1.5585, 1.4885, 1.381, 1.419)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW2    CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17679L, 17680L, 17681L), class = "Date"), px_last = c(1.0335, 
    1.0508, 1.0094, 0.9988, 0.9674, 0.9674, 0.9027, 0.92975)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 8L)), `BPSW10   CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17679L, 17680L, 17681L), class = "Date"), px_last = c(1.651, 
    1.675, 1.616, 1.593, 1.52, 1.52, 1.427, 1.455)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 8L)), `NA` = structure(list(date = structure(integer(0), class = "Date"), 
    px_last = numeric(0)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = integer(0)), `BPSW3    CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.1795, 1.2025, 
    1.1525, 1.1445, 1.0965, 1.01, 1.0435)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW12   CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.7105, 1.7325, 
    1.6735, 1.6495, 1.5795, 1.474, 1.5115)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW4    CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.294, 1.33, 
    1.27, 1.258, 1.202, 1.107, 1.1371)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW15   CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.7615, 1.7805, 
    1.7225, 1.6985, 1.6295, 1.525, 1.5615)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW5    CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.3855, 1.4155, 
    1.3595, 1.3465, 1.2875, 1.185, 1.21425)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW20   CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.7895, 1.8045, 
    1.7485, 1.7255, 1.6565, 1.554, 1.5895)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW6    CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.4565, 1.4855, 
    1.4285, 1.4135, 1.35725, 1.2555, 1.278)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW25   CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.778, 1.791, 
    1.737, 1.716, 1.647, 1.548, 1.5835)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L)), `BPSW7    CMPN Curncy` = structure(list(
    date = structure(c(17672L, 17673L, 17674L, 17675L, 17676L, 
    17680L, 17681L), class = "Date"), px_last = c(1.5155, 1.5445, 
    1.4875, 1.4695, 1.4025, 1.298, 1.331)), class = "data.frame", .Names = c("date", 
"px_last"), row.names = c(NA, 7L))), .Names = c("NA", "BP0012M       Index", 
"BP0003M       Index", "BP0006M       Index", "NA", "NA", "NA", 
"NA", "NA", "BPSW30   CMPN Curncy", "NA", "NA", "BPSW8    CMPN Curncy", 
"BPSW1F   CMPN Curncy", "BPSW9    CMPN Curncy", "BPSW2    CMPN Curncy", 
"BPSW10   CMPN Curncy", "NA", "BPSW3    CMPN Curncy", "BPSW12   CMPN Curncy", 
"BPSW4    CMPN Curncy", "BPSW15   CMPN Curncy", "BPSW5    CMPN Curncy", 
"BPSW20   CMPN Curncy", "BPSW6    CMPN Curncy", "BPSW25   CMPN Curncy", 
"BPSW7    CMPN Curncy"))

【问题讨论】：

你能分享dput(rf_curves$GBP)的输出吗？
> dput(rf_curves$GBP) structure(list(NA = structure(list(date = structure(integer(0), class= "Date"), px_last = numeric(0)) , class= "data.frame", .Names = c("date", "px_last"), row.names = integer(0)), BP0012M Index = structure(list( date = structure(17672:17676, 1234562) = "Date"), px_last = c(0.92894, 等等，但我达到了字符限制，请参阅 paste.ofcode.org 上的完整输出并粘贴到 URL cdP8xak6yuKFQ5V65zgssb
请修改您的问题以包含来自dput(rf_curves$GBP) 的输出，而不是将其粘贴到 cmets 中。

标签： r

【解决方案1】：

一种方法是对数据集进行行绑定，然后使用tidyr::spread。假设你的数据框列表是dat，那么

library(dplyr)
library(tidyr)
out <- bind_rows(dat, .id = "ticker") %>%
  mutate(ticker = gsub("^([A-Z0-9]+).*$", "\\1", ticker)) %>%
  spread(key = ticker, value = px_last)

gsub 清理ticker 以仅包含代码本身。输出看起来像

out[, 1:6]
#         date BP0003M BP0006M BP0012M BPSW10 BPSW12
# 1 2018-05-21 0.62281 0.74630 0.92894  1.651 1.7105
# 2 2018-05-22 0.62250 0.74323 0.93081  1.675 1.7325
# 3 2018-05-23 0.61900 0.73411 0.91831  1.616 1.6735
# 4 2018-05-24 0.61406 0.73321 0.91820  1.593 1.6495
# 5 2018-05-25 0.61067 0.72312 0.90056  1.520 1.5795
# 6 2018-05-28      NA      NA      NA  1.520     NA
# 7 2018-05-29      NA      NA      NA  1.427 1.4740
# 8 2018-05-30      NA      NA      NA  1.455 1.5115

【讨论】：

有效！ \\1 是否选择第一行（即名称）？我不明白 gsub 部分以及如何首先使用 id “ticker”，然后使用 mutate 创建变量？
每个代码都代表一个到期时间 (TTM)。对于 BP0003M，TTM 为 3/12。对于 BP0012M，TTM 为 1。对于 BPSW10，TTM 为 10。对于 BPSW12，TTM 为 12。稍后我将不得不在例如 4y 和 5y TTM 之间进行插值。那么：我可以重命名这些列，然后按照 TTM 的顺序重新排列它们，以便以后更容易插值吗？
如果您有后续问题，我建议为他们打开新问题。 cmets 部分不是提供答案的好地方。此外，提出新问题可以让其他有相同问题的人找到解决方案。