R-从嵌套列表中结合数据帧

Question

我在将包含在嵌套列表中的数据帧与简洁语法组合在一起时遇到了麻烦。我有以下表格的嵌套列表：

nestedList <- lapply(1:3,function(y){
  lapply(1:8,function(z){
    data.frame(matrix(rnorm(20), nrow=10))
  })
})

因此，nestedList包含3个列表，每个列表包含8个带有数据框的列表。我想将列表组合在一起：

tmp1 <- nestedList[[1]][[1]]
tmp2 <- nestedList[[2]][[1]]
tmp3 <- nestedList[[3]][[1]]
expectedResult <- rbind(tmp1,tmp2,tmp3)

我本来希望以下语法有效，但显然不是：

unexpectedResult <- rbind(nestedList[[1:3]][[1]])

Answer 1

尝试一下。

foo <- lapply(nestedList, function(x) x[[1]])
this <- do.call("rbind", foo)

Answer 2

我想出了使用purrr

的以下解决方案

my_result <- nestedList %>%
  # extract first dataframe from each nested list
  map(`[[`, 1) %>% 
  # bind rows together
  bind_rows()

并测试结果正确

identical(my_result, expectedResult)
[1] TRUE

Answer 3

do.call(rbind, lapply(nestedList[1:3], '[[', 1))

将做到这一点：

set.seed(123)
nestedList <- lapply(1:5,function(y){
  lapply(1:8,function(z){
    data.frame(matrix(rnorm(20), nrow=10))
  })
})
> do.call(rbind, lapply(nestedList[1:3], '[[', 1))
            X1          X2
1  -0.56047565  1.22408180
2  -0.23017749  0.35981383
3   1.55870831  0.40077145
4   0.07050839  0.11068272
5   0.12928774 -0.55584113
6   1.71506499  1.78691314
7   0.46091621  0.49785048
8  -1.26506123 -1.96661716
9  -0.68685285  0.70135590
10 -0.44566197 -0.47279141
11  1.05271147 -0.21538051
12 -1.04917701  0.06529303
13 -1.26015524 -0.03406725
14  3.24103993  2.12845190
15 -0.41685759 -0.74133610
16  0.29822759 -1.09599627
17  0.63656967  0.03778840
18 -0.48378063  0.31048075
19  0.51686204  0.43652348
20  0.36896453 -0.45836533
21  0.23743027  1.01755864
22  1.21810861 -1.18843404
23 -1.33877429 -0.72160444
24  0.66082030  1.51921771
25 -0.52291238  0.37738797
26  0.68374552 -2.05222282
27 -0.06082195 -1.36403745
28  0.63296071 -0.20078102
29  1.33551762  0.86577940
30  0.00729009 -0.10188326

Answer 4

我想指出data.table的rbindlist函数。此功能通常比基本rbind

更有效

 library(data.table)
rbindlist(unlist(nestedList, recursive = F))
# Performance comparison
microbenchmark(times = 1000,
   datatable_rbind = rbindlist(unlist(nestedList, recursive = F)),
   base_rbind = do.call("rbind", lapply(nestedList, function(x) x[[1]])),
   base_rbind2 = do.call(rbind, lapply(nestedList[1:3], '[[', 1))
)
# Unit: microseconds
# expr     min      lq     mean   median       uq      max neval
# datatable_rbind  85.530 109.397 124.5534 124.3035 141.1110 216.816  1000
# base_rbind 135.037 152.035 190.5976 184.8475 201.0455 5912.946 1000
# base_rbind2 136.196 151.783 179.9393 186.4245 200.4225  347.564 1000

Answer 5

如果嵌套的数据帧列表具有相同的维度（相同的行名和列名），则可以简单地使用：

merged_df = docall("rbind", nestedList)