按名称合并内部嵌套列表元素

Question

我已经问过这个问题了，但没有很好的表现，所以我决定从头再问一次。

我在R中有一个大List对象，包含100个不同模型的输出列表，每个输出本身就是一个包含模型所有参数的统计信息的列表。因此，我有一个子列表的手段，一个为sd，等等，在每一个100输出列表。每个统计数据对于模型的每个参数都有一个向量或矩阵。我想要的是将所有100个列表输出合并到一个大列表中，方法是根据统计数据和参数名称对它们进行聚合。

modeloutputs <- list("model1"=list("simlist"=list("N"=c(1,2,3,4,5,6,7), "sigma"=c(1,2,3,4), "theta"=c(7,8,9)),
                                   "sd" = list("N"=c(6), "sigma"=c(6.5), "theta"=matrix(6:11, nrow = 2, ncol = 3)), 
                                   "mean" = list("N"=c(3), "sigma"=c(3.5), "theta"=matrix(6:11, nrow = 2, ncol = 3))),
                     "model2"=list("simlist"=list("N"=c(8,9,10,11,12,13,14), "sigma"=c(1,2,3,4), "theta"=c(7,8,9)),
                                   "sd" = list("N"=c(7), "sigma"=c(8.5), "theta"=matrix(6:11, nrow = 2, ncol = 3)), 
                                   "mean" = list("N"=c(5), "sigma"=c(5.5), "theta"=matrix(6:11, nrow = 2, ncol = 3))),
                     "model3"=list("simlist"=list("N"=c(15,16,17,18,19,20,21), "sigma"=c(1,2,3,4), "theta"=c(7,8,9)),
                                   "sd" = list("N"=c(8), "sigma"=c(8.5), "theta"=matrix(6:11, nrow = 2, ncol = 3)), 
                                   "mean" = list("N"=c(9), "sigma"=c(9.5), "theta"=matrix(6:11, nrow = 2, ncol = 3))))


modeloutputs_wanted <- list("allmodels"=list("simlist"=list("N"=c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21), "sigma"=c(1,2,3,4,1,2,3,4,1,2,3,4), "theta"=c(7,8,9,7,8,9,7,8,9)),
                                   "sd" = list("N"=c(6,7,8), "sigma"=c(6.5,8.5,8.5), "theta"=matrix(6:11, nrow = 2, ncol = 3)), 
                                   "mean" = list("N"=c(3,5,9), "sigma"=c(3.5,5.5,9.5), "theta"=matrix(6:11, nrow = 2, ncol = 3))))

theta应该将所有矩阵按行连接起来，就像rowbind()一样，我无法在示例中正确地再现它。

Answer 1

这和你要找的东西很接近：

library(purrr)
transpose(modeloutputs) |>
  map(~ transpose(.x)) %>% 
  map_depth(2, ~ if(all(sapply(.x, is.matrix))) unique(.x) else reduce(.x, c))

输出

$simlist
$simlist$N
 [1]  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20
[21] 21

$simlist$sigma
 [1] 1 2 3 4 1 2 3 4 1 2 3 4

$simlist$theta
[1] 7 8 9 7 8 9 7 8 9


$sd
$sd$N
[1] 6 7 8

$sd$sigma
[1] 6.5 8.5 8.5

$sd$theta
$sd$theta[[1]]
     [,1] [,2] [,3]
[1,]    6    8   10
[2,]    7    9   11



$mean
$mean$N
[1] 3 5 9

$mean$sigma
[1] 3.5 5.5 9.5

$mean$theta
$mean$theta[[1]]
     [,1] [,2] [,3]
[1,]    6    8   10
[2,]    7    9   11

Answer 2

或在R基：

agg_list <- list()
for(statistic in lapply(modeloutputs, names)[[1]]){
  theta_vec<- c()
  theta <- matrix(ncol=3)
  N_vec<- c()
  sigma_vec<- c()
  for(model in names(modeloutputs)){
    if(statistic == "simlist"){
      theta_vec <- c(theta_vec, modeloutputs[[model]][[statistic]][["theta"]])
      sigma_vec <- c(sigma_vec, modeloutputs[[model]][[statistic]][["sigma"]])
      N_vec <- c(N_vec, modeloutputs[[model]][[statistic]][["N"]])
      stats <-list("N"=N_vec, "sigma"=sigma_vec, "theta"=theta_vec )
    }else{
      theta <- rbind(theta, modeloutputs[[model]][[statistic]][["theta"]])
      theta <- na.omit(theta)
      sigma_vec <- c(sigma_vec, modeloutputs[[model]][[statistic]][["sigma"]])
      N_vec <- c(N_vec, modeloutputs[[model]][[statistic]][["N"]])
      stats <- list("N"=N_vec, "sigma"=sigma_vec, "theta"=theta )
    }
  }
  agg_list <- append(agg_list , list(stats))
}
names(agg_list) <- sapply(modeloutputs, names)[,1]

但无可否认，@Ma l的回答要优雅得多:)