R - 替换嵌套 - 嵌套列表

时间:2018-03-08 09:02:50

标签: r list for-loop

我在R中构建一个列表列表,我使用它的唯一方法是使用嵌套的fors。问题是它们需要4秒钟才能运行,因此速度极慢。这就是我构建它们的方式:

tt = vector("list", length(dp$id))

for (idx in seq_along(dp$id)) {
   pos       = dp$id[idx]
   position  = dp[id == pos]$pdim
   tt[[idx]] = list(
      a = unbox(pos),
      b = list()
   )
   tmp = positionData[positionId == position]
   tls = vector("list", nrow(tmp))
   if (nrow(tmp)) {
      for (row in 1:nrow(tmp)) {
         tls[[row]] = list(
            c = unbox(tmp[row]$d),
            d = unbox(tmp[row]$c)
         )
      }
      tt[[idx]]$b = tls
   }
}

有没有办法替换两者以更快地构建列表?

编辑:样本数据

dp = data.table(id =c(5632,5633, 5634, 5635, 5636), pdim = c(2103, 2048, 2093, 2069, 2086))

positionData = data.table(
positionId = c(2048, 2069, 2086, 2093, 2103, 2048, 2069, 2086, 2093, 2103, 2048, 2069, 2086, 2093, 2103, 2048, 2069, 2086, 2093, 2103, 2048, 2069, 2086, 2093, 2103, 2048),

d = c(0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0),
c = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0))

1 个答案:

答案 0 :(得分:1)

每个tt子列表的b(总共5个)是data.frame而不是列表?如果你能忍受这个,你肯定可以避免第二次循环:

library(dplyr)
tt = list()
for (idx in seq_along(dp$id)) {
  pos = dp$id[idx]
  position= dp[id == pos]$pdim
  tt[[idx]] = list(a = pos,b = list())
  tmp = positionData[positionId == position]
  tls<-transmute(tmp,c=d,d=c)
  tt[[idx]]$b = tls
}