Question

我有一个数据框列表，每个数据框中都有一些重叠的列。列表中的数据帧数量未知。我怎样才能有效地在基础上将数据帧组合在一起并用零填充非重叠列？

示例数据：

x <- data.frame(a=1:2, b=1:2, c=1:2)
y <- data.frame(a=1:2, r=1:2, f=1:2)
z <- data.frame(b=1:3, c=1:3, v=1:3, t=c("A", "A", "D"))

L1 <- list(x, y, z)

期望的输出：

  a b c f r t v
1 1 1 1 0 0 0 0
2 2 2 2 0 0 0 0
3 1 0 0 1 1 0 0
4 2 0 0 2 2 0 0
5 0 1 1 0 0 A 1
6 0 2 2 0 0 A 2
7 0 3 3 0 0 D 3

Answer 1

用缺少的列填充每个数据框，然后对它们进行rbind：

allnames <- unique(unlist(lapply(L1, names)))
do.call(rbind, lapply(L1, function(df) {
    not <- allnames[!allnames %in% names(df)]
    df[, not] <- 0
    df
}))

Answer 2

我有an old (and probably inefficient) function这样做。我在这里做了一个修改，允许指定填充。

RBIND <- function(datalist, keep.rownames = TRUE, fill = NA) {
  Len <- sapply(datalist, ncol)
  if (all(diff(Len) == 0)) {
    temp <- names(datalist[[1]])
    if (all(sapply(datalist, function(x) names(x) %in% temp))) tryme <- "basic"
    else tryme <- "complex"
  } 
  else tryme <- "complex"
  almost <- switch(
    tryme,
    basic = { do.call("rbind", datalist) },
    complex = {
      Names <- unique(unlist(lapply(datalist, names)))
      NROWS <- c(0, cumsum(sapply(datalist, nrow)))
      NROWS <- paste(NROWS[-length(NROWS)]+1, NROWS[-1], sep=":")
      out <- lapply(1:length(datalist), function(x) {
        emptyMat <- matrix(fill, nrow = nrow(datalist[[x]]), ncol = length(Names))
        colnames(emptyMat) <- Names
        emptyMat[, match(names(datalist[[x]]), 
                         colnames(emptyMat))] <- as.matrix(datalist[[x]])
        emptyMat
      })
      do.call("rbind", out)
    })
  Final <- as.data.frame(almost, row.names = 1:nrow(almost))
  Final <- data.frame(lapply(Final, function(x) type.convert(as.character(x))))
  if (isTRUE(keep.rownames)) {
    row.names(Final) <- make.unique(unlist(lapply(datalist, row.names)))
  } 
  Final
}

这是您的样本数据。

RBIND(L1, fill = 0)
#     a b c r f v t
# 1   1 1 1 0 0 0 0
# 2   2 2 2 0 0 0 0
# 1.1 1 0 0 1 1 0 0
# 2.1 2 0 0 2 2 0 0
# 1.2 0 1 1 0 0 1 A
# 2.2 0 2 2 0 0 2 A
# 3   0 3 3 0 0 3 D

任意数量的数据帧的rbind列表

2 个答案: