從R中的多個資料框中過濾行

我有五個data frame尺寸相同（8 x 2）但不同column name的 s，如下所示：

nbb <- (
nbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
nbb_RMSE = c(1.0152338, 0.7199394, 0.7990978, 0.9045563, 1.6514406, 0.5160516, 0.4964024, 0.2617795)
)

mbb <- (
mbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
mbb_RMSE = c(0.8324074, 0.9278236, 1.9817984, 0.9567368, 0.2814623, 0.1129459, 0.1233126, 0.4222578)
)

cbb <- (
cbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
cbb_RMSE = c(1.27782499, 1.96332220, 0.74704997, 0.46579943, 1.10850563, 0.40456698, 0.26027359, 0.02452239)
)

tmbb <- (
tmbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
tmbb_RMSE = c(0.83240742, 1.05126826, 0.08290467, 0.76397988, 1.23772208, 0.57628337, 0.56437185, 0.46460279)
)

tcbb <- (
tcbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
tcbb_RMSE = c(0.9328510, 0.8312332, 0.9402116, 1.6029357, 2.0001519, 0.4387557, 0.5965582, 0.4148854)
)

我想創建一個新的，它將包含上面五（5）sdata frame中包含的每一行。如果我將不同的 s 命名為 (nbb, mbb, cbb, tmbb, tcbb)minimum RMSEdata framerow namesdata frame

rownames(df) <- c("nbb", "mbb", "cbb", "tmbb", "tcbb")

我希望得到如下結果：

uj5u.com熱心網友回復：

首先將你的 5 個資料框組合成一個串列，然后使用它lapply來遍歷所有資料框并輸出必要的資訊。lapply()此外，使用do.callto row bind ( rbind) 結果包裝。最后，將結果更改為資料框并將行名轉換為列名

rownames_to_column("df")如果要使用行名稱而不是存盤資料框名稱的列，則可以跳過。

library(tibble)

df_list <- list(nbb = nbb, mbb = mbb, cbb = cbb, tmbb = tmbb, tcbb = tcbb)

do.call(rbind, lapply(df_list, function(x) data.frame(lb = x[which.min(x[,2]), 1], RMSE = min(x[, 2])))) %>%
  rownames_to_column("df")

    df lb       RMSE
1  nbb  9 0.26177950
2  mbb  7 0.11294590
3  cbb  9 0.02452239
4 tmbb  4 0.08290467
5 tcbb  9 0.41488540

輸入資料

此外，請注意您在帖子中定義資料框的方式不正確，我在此處包含了“正確”的方式以供參考。

nbb <- data.frame(
  nbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  nbb_RMSE = c(1.0152338, 0.7199394, 0.7990978, 0.9045563, 1.6514406, 0.5160516, 0.4964024, 0.2617795))

mbb <- data.frame( mbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  mbb_RMSE = c(0.8324074, 0.9278236, 1.9817984, 0.9567368, 0.2814623, 0.1129459, 0.1233126, 0.4222578)
)

cbb <- data.frame(
  cbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  cbb_RMSE = c(1.27782499, 1.96332220, 0.74704997, 0.46579943, 1.10850563, 0.40456698, 0.26027359, 0.02452239)
)

tmbb <- data.frame(
  tmbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  tmbb_RMSE = c(0.83240742, 1.05126826, 0.08290467, 0.76397988, 1.23772208, 0.57628337, 0.56437185, 0.46460279))

tcbb <- data.frame(
  tcbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  tcbb_RMSE = c(0.9328510, 0.8312332, 0.9402116, 1.6029357, 2.0001519, 0.4387557, 0.5965582, 0.4148854)
  )

uj5u.com熱心網友回復：

如果您愿意接受一種purrr方法，您可以使用

library(purrr)
library(dplyr)

my_list %>% 
  map_dfr(~.x %>% 
            filter(if_any(ends_with("_RMSE"), ~.x == min(.x))) %>% 
            rename_with(~gsub(".*_", "", .x)), 
          .id = "df") %>% 
  bind_rows()

這回傳

    df lb       RMSE
1  cbb  9 0.02452239
2  mbb  7 0.11294590
3  nbb  9 0.26177950
4 tcbb  9 0.41488540
5 tmbb  4 0.08290467

with (借自 Rui Barradas)

my_list <- mget(ls(pattern = "bb$"))

uj5u.com熱心網友回復：

這是一個基本的 R 方式。
首先將 data.frames 放在一個帶有 的串列中mget，然后lapply串列是一個匿名函式，輸出最小的 RMSE 行。系結輸出行并根據想要的行名重新排序。

nbb <- data.frame(
  nbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  nbb_RMSE = c(1.0152338, 0.7199394, 0.7990978, 0.9045563, 1.6514406, 0.5160516, 0.4964024, 0.2617795)
)

mbb <- data.frame(
  mbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  mbb_RMSE = c(0.8324074, 0.9278236, 1.9817984, 0.9567368, 0.2814623, 0.1129459, 0.1233126, 0.4222578)
)

cbb <- data.frame(
  cbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  cbb_RMSE = c(1.27782499, 1.96332220, 0.74704997, 0.46579943, 1.10850563, 0.40456698, 0.26027359, 0.02452239)
)

tmbb <- data.frame(
  tmbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  tmbb_RMSE = c(0.83240742, 1.05126826, 0.08290467, 0.76397988, 1.23772208, 0.57628337, 0.56437185, 0.46460279)
)

tcbb <- data.frame(
  tcbb_lb = c(2, 3, 4, 5, 6, 7, 8, 9),
  tcbb_RMSE = c(0.9328510, 0.8312332, 0.9402116, 1.6029357, 2.0001519, 0.4387557, 0.5965582, 0.4148854)
)

df_list <- mget(ls(pattern = "bb$"))

tmp <- lapply(df_list, \(x){
  i <- which.min(x[[2]])
  if(length(i) > 0L) {
    data.frame(lb = x[i, 1], RMSE = x[i, 2])
  } else NULL
})
res <- do.call(rbind, tmp)
rm(tmp)
res <- cbind.data.frame(df = names(df_list), res)

i <- order(c("nbb", "mbb", "cbb", "tmbb", "tcbb"))
res <- res[i,]
res
#>        df lb       RMSE
#> nbb   nbb  9 0.26177950
#> mbb   mbb  7 0.11294590
#> cbb   cbb  9 0.02452239
#> tmbb tmbb  4 0.08290467
#> tcbb tcbb  9 0.41488540

^{由reprex 包于 2022-04-10 創建(v2.0.1)}

標籤：r 数据框 筛选

上一篇：如何在.netcore中使用Session？

下一篇：從熊貓資料框中選擇隨機值并向它們添加常量