在r中重新格式化数据

时间:2012-08-16 13:54:46

标签: r dataframe reorganize

我有以下类型的数据,尽管真实数据集中的set和indvid级别相当高:

set <- c(rep(1,6), rep(2,6))
Indvidual <- c(rep (c("IndvA", "IndvA", "IndvB", "IndvB", "IndvC", "IndvC"), 2))
leftposition <- c(10,  10,0 ,0,  0,   0,  40,     40,    30,  30,  20,  20 )
rightposition <- c(20,  20,20,20,  30,  30, 50,     50,    40,  40,  60,  60 )
leftmark <- c(     1, 3, 5,    7,    9,   11,  13,       15,     17,   19,   21 , 23 )
rightmark <- c( 2,    4,     6,    8,    10,   12,14,      16,  18,   20,   22,  24 )

myd <- data.frame (set, Indvidual,leftposition,rightposition, leftmark, rightmark)
myd

      set Indvidual leftposition rightposition leftmark rightmark
1    1     IndvA           10            20        1         2
2    1     IndvA           10            20        3         4
3    1     IndvB            0            20        5         6
4    1     IndvB            0            20        7         8
5    1     IndvC            0            30        9        10
6    1     IndvC            0            30       11        12
7    2     IndvA           40            50       13        14
8    2     IndvA           40            50       15        16
9    2     IndvB           30            40       17        18
10   2     IndvB           30            40       19        20
11   2     IndvC           20            60       21        22
12   2     IndvC           20            60       23        24

在第一列为Individual的新数据集中,其余列将全部唯一(leftpostion,rightposition)

sort (unique (c(leftposition, rightposition)))
[1]  0 10 20 30 40 50 60

现在对于set = 1,我想为个人添加值(注意每个Indvidual已重复两次,是预期的)。每个人都有两个值 - 一个添加到左侧(左侧)另一个添加到右侧(右侧位置)。要向左或向右打印的实际数据分别在左标记和右标记中。因此,对于第一组,有组织的数据将如下所示:

enter image description here

然后将set2(或n设置数据)添加到同一个表中。最后的任何空白都将填充NA或指定的任何值(例如“ - ”)。

enter image description here

感谢您的帮助:

1 个答案:

答案 0 :(得分:3)

library(reshape2)
library(plyr)
#Make indviduals unique
myd <- ddply(myd, .(set, Indvidual), transform, 
             Indvidual = paste(Indvidual, order(Indvidual), sep = "_"))

# bind positions together

myd_molten <- melt(myd, id.vars=c("set", "Indvidual"))
marks <- grep("mark", myd_molten$variable)
levels(myd_molten$variable) <- rep(c("left", "right"), 2)
myd_pos <- myd_molten[-marks,]
names(myd_pos)[4] <- "position"
myd_mark <- myd_molten[marks,]
myd_binded <- cbind(myd_pos, mark = myd_mark$value)

#cast it into the desired form and get the names right
#I could have done the names with gsub but I didn't want to mess with regexpr

ans <- dcast(Indvidual ~ position, value.var = "mark", data = myd_binded)
ans$Indvidual <- do.call(rbind, strsplit(ans$Indvidual, "_"))[,1]
ans

  Indvidual  0 10 20 30 40 50 60
1     IndvA NA  1  2 NA 13 14 NA
2     IndvA NA  3  4 NA 15 16 NA
3     IndvB  5 NA  6 17 18 NA NA
4     IndvB  7 NA  8 19 20 NA NA
5     IndvC  9 NA 21 10 NA NA 22
6     IndvC 11 NA 23 12 NA NA 24