Question

我有一个大数据集，下面给出了一个样本：

df <- data.frame(stringsAsFactors=FALSE,
                 Date = c("2015-10-26", "2015-10-26", "2015-10-26", "2015-10-26",
                          "2015-10-27", "2015-10-27", "2015-10-27"),
                 Ticker = c("ANZ", "CBA", "NAB", "WBC", "ANZ", "CBA", "WBC"),
                 Open = c(29.11, 77.89, 32.69, 31.87, 29.05, 77.61, 31.84),
                 High = c(29.17, 77.93, 32.76, 31.92, 29.08, 78.1, 31.95),
                 Low = c(28.89, 77.37, 32.42, 31.71, 28.9, 77.54, 31.65),
                 Close = c(28.9, 77.5, 32.42, 31.84, 28.94, 77.74, 31.77),
                 Volume = c(6350170L, 2251288L, 3804239L, 5597684L, 5925519L, 2424679L,
                            5448863L)
)

我想解决的问题是NAB在2015年10月27日缺少的数据

我希望最后一个值为缺少的日期重复一遍：

    Date Ticker  Open  High   Low Close  Volume

2 2015-10-27    NAB 32.69 32.76 32.42 32.42 3804239

关于如何做到这一点的任何想法？

我已尝试gather + spread

但未成功

Answer 1

如果你尝试过类似的东西怎么办？

LOCF

我假设如果不存在Ticker / Day组合，你想创建一个和expand.grid它。这就是{{1}}的作用。

Answer 2

tidyr::complete和tidyr::fill仅针对这种情况构建：

library(tidyverse)
df %>%
  complete(Date,Ticker) %>%
  arrange(Ticker) %>%
  fill(names(.)) %>%
  arrange(Date)
# 
# # A tibble: 8 x 7
#         Date Ticker  Open  High   Low Close  Volume
#        <chr>  <chr> <dbl> <dbl> <dbl> <dbl>   <int>
# 1 2015-10-26    ANZ 29.11 29.17 28.89 28.90 6350170
# 2 2015-10-26    CBA 77.89 77.93 77.37 77.50 2251288
# 3 2015-10-26    NAB 32.69 32.76 32.42 32.42 3804239
# 4 2015-10-26    WBC 31.87 31.92 31.71 31.84 5597684
# 5 2015-10-27    ANZ 29.05 29.08 28.90 28.94 5925519
# 6 2015-10-27    CBA 77.61 78.10 77.54 77.74 2424679
# 7 2015-10-27    NAB 32.69 32.76 32.42 32.42 3804239
# 8 2015-10-27    WBC 31.84 31.95 31.65 31.77 5448863

Answer 3

另一个可能的解决方案（注意：我必须将您的日期向量转换为日期格式，但这可以在最终输出中反转）：

library(tidyr)
library(dplyr)

df <- data.frame(stringsAsFactors=FALSE,
                 Date = as.Date(c("2015-10-26", "2015-10-26", "2015-10-26", "2015-10-26",
                          "2015-10-27", "2015-10-27", "2015-10-27")),
                 Ticker = c("ANZ", "CBA", "NAB", "WBC", "ANZ", "CBA", "WBC"),
                 Open = c(29.11, 77.89, 32.69, 31.87, 29.05, 77.61, 31.84),
                 High = c(29.17, 77.93, 32.76, 31.92, 29.08, 78.1, 31.95),
                 Low = c(28.89, 77.37, 32.42, 31.71, 28.9, 77.54, 31.65),
                 Close = c(28.9, 77.5, 32.42, 31.84, 28.94, 77.74, 31.77),
                 Volume = c(6350170L, 2251288L, 3804239L, 5597684L, 5925519L, 2424679L,
                            5448863L))
tickers<- unique(df$Ticker)               
dates<- as.Date(df$Date)

possibilities<- as.data.frame(unique(expand.grid(dates,tickers)))
colnames(possibilities)<- c('Date','Ticker')

missing<- anti_join(possibilities,df[,c('Date','Ticker')])

missing_filled<- if(nrow(missing)>0){
replacement<-   cbind(missing,filter(df,Date==missing$Date-1,Ticker==missing$Ticker)[,3:7])
}

final<- arrange(rbind(df,replacement),Date)

填写缺失的行

3 个答案: