在 R 的数据框中计算具有非连续天数的变量的滚动总和
Work out rolling sums for variables with non-consecutive days in a dataframe in R
我有一些数据，我想根据这些数据计算出连续 14 天 window 的获胜百分比，尽管有大约 7 年的结果。这些日子是不连续的，所以每当我按 'Trainer' 变量和 运行 rollapplyr 或 runSum/ sum_run 分组时，我都会总结过去的 14 个事件，但是无法弄清楚如何将 14 天分组。当我尝试使用日期中的日期定义宽度或 k 值时，出现错误
invalid time series parameters specified
或vec' must be sorted non-decreasingly and not contain NAs
编辑- 下面的代码给出了上面的错误
df %>% group_by(Trainer) %>% mutate(Fourteen_day_wins = rollapplyr(Wins, width = 1:n() - findInterval( Date %d-% Days(14), Date), sum)) %>% ungroup
我想在我的 df 的新列中按 Trainer 分组时获得 14 天滚动期间的总获胜次数和事件数。有人可以指出我正确的方向吗？仍然是 R 的新手所以到目前为止一直难倒我！
样本 df:
structure(list(Trainer = c("Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J"), Wins = c(1, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0), Date = structure(c(1508025600, 1508112000, 1508112000, 1508112000, 1508198400, 1508284800, 1508284800, 1508284800, 1508457600, 1508457600, 1508544000, 1508544000, 1508544000, 1508716800, 1508716800, 1508716800, 1508803200, 1508803200, 1508803200, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1509062400, 1509062400, 1509062400, 1509062400, 1509062400, 1509148800, 1509148800, 1509148800, 1509148800, 1509148800, 1509148800, 1509321600, 1509321600, 1509321600, 1509321600, 1509494400, 1509667200, 1509667200, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1510099200, 1510099200, 1510099200, 1510358400, 1510358400, 1510358400, 1521936000, 1521936000, 1523923200, 1523923200, 1523923200, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524182400, 1524182400, 1524182400, 1524268800, 1524268800, 1524268800, 1524528000, 1524528000, 1524528000, 1524528000, 1524614400, 1524614400, 1524614400, 1524787200, 1524787200, 1524787200, 1524787200, 1524787200, 1525132800, 1525219200, 1525219200, 1525219200), tzone = "UTC", class = c("POSIXct", "POSIXt"))), row.names = c(NA, -101L), class = c("tbl_df", "tbl", "data.frame"))
一种选择是将所有天数和所有培训师合并，将其与原始数据合并，然后使用 14 天 window:
library(zoo)
#>
#> Attaching package: 'zoo'
#> The following objects are masked from 'package:base':
#>
#> as.Date, as.Date.numeric
library(tidyverse)
df <- structure(list(Trainer = c("Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J"), Wins = c(1, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0), Date = structure(c(1508025600, 1508112000, 1508112000, 1508112000, 1508198400, 1508284800, 1508284800, 1508284800, 1508457600, 1508457600, 1508544000, 1508544000, 1508544000, 1508716800, 1508716800, 1508716800, 1508803200, 1508803200, 1508803200, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1509062400, 1509062400, 1509062400, 1509062400, 1509062400, 1509148800, 1509148800, 1509148800, 1509148800, 1509148800, 1509148800, 1509321600, 1509321600, 1509321600, 1509321600, 1509494400, 1509667200, 1509667200, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1510099200, 1510099200, 1510099200, 1510358400, 1510358400, 1510358400, 1521936000, 1521936000, 1523923200, 1523923200, 1523923200, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524182400, 1524182400, 1524182400, 1524268800, 1524268800, 1524268800, 1524528000, 1524528000, 1524528000, 1524528000, 1524614400, 1524614400, 1524614400, 1524787200, 1524787200, 1524787200, 1524787200, 1524787200, 1525132800, 1525219200, 1525219200, 1525219200), tzone = "UTC", class = c("POSIXct", "POSIXt"))), row.names = c(NA, -101L), class = c("tbl_df", "tbl", "data.frame"))
all_dates <- with(df, expand_grid(Trainer = unique(Trainer),
Date = seq(min(Date), max(Date), by="1 day")))
all_dates <- left_join(all_dates, df)
#> Joining, by = c("Trainer", "Date")
all_dates %>%
group_by(Trainer) %>%
mutate(win_pct = rollapplyr(Wins,
width=14,
mean,
partial = TRUE,
align="right",
na.rm=TRUE,
fill=TRUE))
#> # A tibble: 460 × 4
#> # Groups: Trainer [2]
#> Trainer Date Wins win_pct
#> <chr> <dttm> <dbl> <dbl>
#> 1 Appleby, Charlie 2017-10-15 00:00:00 1 1
#> 2 Appleby, Charlie 2017-10-16 00:00:00 NA 1
#> 3 Appleby, Charlie 2017-10-17 00:00:00 NA 1
#> 4 Appleby, Charlie 2017-10-18 00:00:00 1 1
#> 5 Appleby, Charlie 2017-10-18 00:00:00 0 0.667
#> 6 Appleby, Charlie 2017-10-19 00:00:00 NA 0.667
#> 7 Appleby, Charlie 2017-10-20 00:00:00 NA 0.667
#> 8 Appleby, Charlie 2017-10-21 00:00:00 NA 0.667
#> 9 Appleby, Charlie 2017-10-22 00:00:00 NA 0.667
#> 10 Appleby, Charlie 2017-10-23 00:00:00 0 0.5
#> # … with 450 more rows
由 reprex package (v2.0.1)
创建于 2022-05-31
问题是 findInterval 的参数应该是数字并且是有序的。
为了解决这个问题，将日期转换为日期 class，然后转换为数字，这样下面的 d 就是自纪元以来的天数。现在我们可以将它与 findInterval 一起使用，如图所示。如果数据已经排序，则可以省略排列行。
library(dplyr, exclude = c("filter", "lag"))
library(zoo)
DF %>%
arrange(Trainer, Date) %>%
group_by(Trainer) %>%
mutate(d = as.numeric(as.Date(Date)),
Wins14 = rollapplyr(Wins, 1:n() - findInterval(d - 14, d), sum)) %>%
ungroup %>%
select(-d)
给予：
# A tibble: 101 x 4
Trainer Wins Date Wins14
<chr> <dbl> <dttm> <dbl>
1 Appleby, Charlie 1 2017-10-15 00:00:00 1
2 Appleby, Charlie 1 2017-10-18 00:00:00 2
3 Appleby, Charlie 0 2017-10-18 00:00:00 2
4 Appleby, Charlie 0 2017-10-23 00:00:00 2
5 Appleby, Charlie 1 2017-10-25 00:00:00 3
6 Appleby, Charlie 0 2017-10-25 00:00:00 3
7 Appleby, Charlie 0 2017-10-25 00:00:00 3
8 Appleby, Charlie 1 2017-10-25 00:00:00 4
9 Appleby, Charlie 0 2017-10-27 00:00:00 4
10 Appleby, Charlie 0 2017-10-27 00:00:00 4
# ... with 91 more rows
您可以使用 complete 来完成您的数据，然后使用 14 window 句点
df %>%
group_by(Trainer) %>%
complete(Date = seq(min(Date), max(Date), '1 day')) %>%
mutate(runMeans = zoo::rollmean(Wins, 14,0,na.rm = TRUE))
# A tibble: 459 x 4
# Groups: Trainer [2]
Trainer Date Wins runMeans
<chr> <dttm> <dbl> <dbl>
1 Appleby, Charlie 2017-10-15 00:00:00 1 0
2 Appleby, Charlie 2017-10-16 00:00:00 NA 0
3 Appleby, Charlie 2017-10-17 00:00:00 NA 0
4 Appleby, Charlie 2017-10-18 00:00:00 1 0
5 Appleby, Charlie 2017-10-18 00:00:00 0 0
6 Appleby, Charlie 2017-10-19 00:00:00 NA 0
7 Appleby, Charlie 2017-10-20 00:00:00 NA 0.429
8 Appleby, Charlie 2017-10-21 00:00:00 NA 0.429
9 Appleby, Charlie 2017-10-22 00:00:00 NA 0.429
10 Appleby, Charlie 2017-10-23 00:00:00 0 0.375
我有一些数据，我想根据这些数据计算出连续 14 天 window 的获胜百分比，尽管有大约 7 年的结果。这些日子是不连续的，所以每当我按 'Trainer' 变量和 运行 rollapplyr 或 runSum/ sum_run 分组时，我都会总结过去的 14 个事件，但是无法弄清楚如何将 14 天分组。当我尝试使用日期中的日期定义宽度或 k 值时，出现错误
invalid time series parameters specified
或vec' must be sorted non-decreasingly and not contain NAs
编辑- 下面的代码给出了上面的错误
df %>% group_by(Trainer) %>% mutate(Fourteen_day_wins = rollapplyr(Wins, width = 1:n() - findInterval( Date %d-% Days(14), Date), sum)) %>% ungroup
我想在我的 df 的新列中按 Trainer 分组时获得 14 天滚动期间的总获胜次数和事件数。有人可以指出我正确的方向吗？仍然是 R 的新手所以到目前为止一直难倒我！
样本 df:
structure(list(Trainer = c("Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J"), Wins = c(1, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0), Date = structure(c(1508025600, 1508112000, 1508112000, 1508112000, 1508198400, 1508284800, 1508284800, 1508284800, 1508457600, 1508457600, 1508544000, 1508544000, 1508544000, 1508716800, 1508716800, 1508716800, 1508803200, 1508803200, 1508803200, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1509062400, 1509062400, 1509062400, 1509062400, 1509062400, 1509148800, 1509148800, 1509148800, 1509148800, 1509148800, 1509148800, 1509321600, 1509321600, 1509321600, 1509321600, 1509494400, 1509667200, 1509667200, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1510099200, 1510099200, 1510099200, 1510358400, 1510358400, 1510358400, 1521936000, 1521936000, 1523923200, 1523923200, 1523923200, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524182400, 1524182400, 1524182400, 1524268800, 1524268800, 1524268800, 1524528000, 1524528000, 1524528000, 1524528000, 1524614400, 1524614400, 1524614400, 1524787200, 1524787200, 1524787200, 1524787200, 1524787200, 1525132800, 1525219200, 1525219200, 1525219200), tzone = "UTC", class = c("POSIXct", "POSIXt"))), row.names = c(NA, -101L), class = c("tbl_df", "tbl", "data.frame"))
一种选择是将所有天数和所有培训师合并，将其与原始数据合并，然后使用 14 天 window:
library(zoo)
#>
#> Attaching package: 'zoo'
#> The following objects are masked from 'package:base':
#>
#> as.Date, as.Date.numeric
library(tidyverse)
df <- structure(list(Trainer = c("Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J", "Haggas, W J", "Appleby, Charlie", "Haggas, W J", "Appleby, Charlie", "Appleby, Charlie", "Haggas, W J"), Wins = c(1, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0), Date = structure(c(1508025600, 1508112000, 1508112000, 1508112000, 1508198400, 1508284800, 1508284800, 1508284800, 1508457600, 1508457600, 1508544000, 1508544000, 1508544000, 1508716800, 1508716800, 1508716800, 1508803200, 1508803200, 1508803200, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1508889600, 1509062400, 1509062400, 1509062400, 1509062400, 1509062400, 1509148800, 1509148800, 1509148800, 1509148800, 1509148800, 1509148800, 1509321600, 1509321600, 1509321600, 1509321600, 1509494400, 1509667200, 1509667200, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1509753600, 1510099200, 1510099200, 1510099200, 1510358400, 1510358400, 1510358400, 1521936000, 1521936000, 1523923200, 1523923200, 1523923200, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524009600, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524096000, 1524182400, 1524182400, 1524182400, 1524268800, 1524268800, 1524268800, 1524528000, 1524528000, 1524528000, 1524528000, 1524614400, 1524614400, 1524614400, 1524787200, 1524787200, 1524787200, 1524787200, 1524787200, 1525132800, 1525219200, 1525219200, 1525219200), tzone = "UTC", class = c("POSIXct", "POSIXt"))), row.names = c(NA, -101L), class = c("tbl_df", "tbl", "data.frame"))
all_dates <- with(df, expand_grid(Trainer = unique(Trainer),
Date = seq(min(Date), max(Date), by="1 day")))
all_dates <- left_join(all_dates, df)
#> Joining, by = c("Trainer", "Date")
all_dates %>%
group_by(Trainer) %>%
mutate(win_pct = rollapplyr(Wins,
width=14,
mean,
partial = TRUE,
align="right",
na.rm=TRUE,
fill=TRUE))
#> # A tibble: 460 × 4
#> # Groups: Trainer [2]
#> Trainer Date Wins win_pct
#> <chr> <dttm> <dbl> <dbl>
#> 1 Appleby, Charlie 2017-10-15 00:00:00 1 1
#> 2 Appleby, Charlie 2017-10-16 00:00:00 NA 1
#> 3 Appleby, Charlie 2017-10-17 00:00:00 NA 1
#> 4 Appleby, Charlie 2017-10-18 00:00:00 1 1
#> 5 Appleby, Charlie 2017-10-18 00:00:00 0 0.667
#> 6 Appleby, Charlie 2017-10-19 00:00:00 NA 0.667
#> 7 Appleby, Charlie 2017-10-20 00:00:00 NA 0.667
#> 8 Appleby, Charlie 2017-10-21 00:00:00 NA 0.667
#> 9 Appleby, Charlie 2017-10-22 00:00:00 NA 0.667
#> 10 Appleby, Charlie 2017-10-23 00:00:00 0 0.5
#> # … with 450 more rows
由 reprex package (v2.0.1)
创建于 2022-05-31问题是 findInterval 的参数应该是数字并且是有序的。
为了解决这个问题，将日期转换为日期 class，然后转换为数字，这样下面的 d 就是自纪元以来的天数。现在我们可以将它与 findInterval 一起使用，如图所示。如果数据已经排序，则可以省略排列行。
library(dplyr, exclude = c("filter", "lag"))
library(zoo)
DF %>%
arrange(Trainer, Date) %>%
group_by(Trainer) %>%
mutate(d = as.numeric(as.Date(Date)),
Wins14 = rollapplyr(Wins, 1:n() - findInterval(d - 14, d), sum)) %>%
ungroup %>%
select(-d)
给予：
# A tibble: 101 x 4
Trainer Wins Date Wins14
<chr> <dbl> <dttm> <dbl>
1 Appleby, Charlie 1 2017-10-15 00:00:00 1
2 Appleby, Charlie 1 2017-10-18 00:00:00 2
3 Appleby, Charlie 0 2017-10-18 00:00:00 2
4 Appleby, Charlie 0 2017-10-23 00:00:00 2
5 Appleby, Charlie 1 2017-10-25 00:00:00 3
6 Appleby, Charlie 0 2017-10-25 00:00:00 3
7 Appleby, Charlie 0 2017-10-25 00:00:00 3
8 Appleby, Charlie 1 2017-10-25 00:00:00 4
9 Appleby, Charlie 0 2017-10-27 00:00:00 4
10 Appleby, Charlie 0 2017-10-27 00:00:00 4
# ... with 91 more rows
您可以使用 complete 来完成您的数据，然后使用 14 window 句点
df %>%
group_by(Trainer) %>%
complete(Date = seq(min(Date), max(Date), '1 day')) %>%
mutate(runMeans = zoo::rollmean(Wins, 14,0,na.rm = TRUE))
# A tibble: 459 x 4
# Groups: Trainer [2]
Trainer Date Wins runMeans
<chr> <dttm> <dbl> <dbl>
1 Appleby, Charlie 2017-10-15 00:00:00 1 0
2 Appleby, Charlie 2017-10-16 00:00:00 NA 0
3 Appleby, Charlie 2017-10-17 00:00:00 NA 0
4 Appleby, Charlie 2017-10-18 00:00:00 1 0
5 Appleby, Charlie 2017-10-18 00:00:00 0 0
6 Appleby, Charlie 2017-10-19 00:00:00 NA 0
7 Appleby, Charlie 2017-10-20 00:00:00 NA 0.429
8 Appleby, Charlie 2017-10-21 00:00:00 NA 0.429
9 Appleby, Charlie 2017-10-22 00:00:00 NA 0.429
10 Appleby, Charlie 2017-10-23 00:00:00 0 0.375