问题描述
d1 <- data.frame(id = c("a","b","c"),val = c(111,444,777),month = c( "jan","Jan","jan" ))
d2 <- data.frame(id = c("a",val = c(222,555,888),month = c( "feb","feb","feb" ))
d3 <- data.frame(id = c("a",val = c(333,666,999),month = c( "mar","mar","mar" ))
old.list <- list(d1,d2,d3)
s1 <- data.frame(id = c("a","a","a"),222,333),"mar" ))
s2 <- data.frame(id = c("b","b"),val = c(444,666),month = c( "Jan","mar" ))
s3 <- data.frame(id = c("c","c",val = c(777,888,"mar" ))
new.list <- list(s1,s2,s3)
'old.list' 中的所有数据框大小相同,具有相同的唯一 ID。我正在尝试将“old.list”转换为每个唯一 id 的时间序列,其中“month”是时间戳,即“new.list”,而不使用 for 循环。
补充说明,'old.list'中每个数据框的第n行被组合起来为'new.list'创建新的数据框,提前致谢。
解决方法
在基础 R 中,您可以 rbind
您的列表元素和 split
id:
new.list <- do.call(rbind,old.list)
new.list <- split(new.list,new.list$id)
,
假设您的所有系列都在 1 月开始,这应该可行:
(如果不查看 ts 的 start 参数)
d1 <- data.frame(id = c("a","b","c"),val = c(111,444,777),month = c( "jan","Jan","jan" ))
d2 <- data.frame(id = c("a",val = c(222,555,888),month = c( "feb","feb","feb" ))
d3 <- data.frame(id = c("a",val = c(333,666,999),month = c( "mar","mar","mar" ))
old.list <- list(d1,d2,d3)
d <- Reduce(rbind,old.list)
d <- split(d$val,d$id)
lapply(d,ts,frequency = 12)
#> $a
#> Jan Feb Mar
#> 1 111 222 333
#>
#> $b
#> Jan Feb Mar
#> 1 444 555 666
#>
#> $c
#> Jan Feb Mar
#> 1 777 888 999
由 reprex package (v0.3.0) 于 2021 年 3 月 12 日创建
,这是一个 tidyverse
方法:
d1 <- data.frame(id = c("a",d3)
library(dplyr)
library(purrr)
old.list %>%
reduce(rbind,by = 'id') %>%
filter(id != 'id') %>%
group_split(id)
#> [[1]]
#> # A tibble: 3 x 3
#> id val month
#> <chr> <chr> <chr>
#> 1 a 111 jan
#> 2 a 222 feb
#> 3 a 333 mar
#>
#> [[2]]
#> # A tibble: 3 x 3
#> id val month
#> <chr> <chr> <chr>
#> 1 b 444 Jan
#> 2 b 555 feb
#> 3 b 666 mar
#>
#> [[3]]
#> # A tibble: 3 x 3
#> id val month
#> <chr> <chr> <chr>
#> 1 c 777 jan
#> 2 c 888 feb
#> 3 c 999 mar
由 reprex package (v0.3.0) 于 2021 年 3 月 12 日创建