根据R Program

Question

我的数据

Type1   Type2   Type3       Expected_Output                 
Red     Orange  Pink        Pink                
Green   abc     na          abc             
Blue     na     na          Blue                
white    na     Green       Green               
na     Brown    purple      purple              
na     black     na         black               
grey    na       na         grey                

如何实现以下预期输出：

1. 如果填写了type3，则预期输出应为type3的内容
2. 如果填写了type2且type3为空（na），则预期输出应为type2的内容
3. 如果type3和type2为空（na），则应为expcetd输出 type1的内容

Answer 1

由相同数据类型的多列组成的data.frame表示数据可能从宽格式转换为长格式。因此，melt()用于删除NA和 join 以将新列添加到原始data.frame：

library(data.table)
DT[melt(DT[, rn := .I], id.vars = "rn", na.rm = TRUE)[
    order(variable), .(New = last(value)), by = rn], on = .(rn)][, rn := NULL][]

   Type1  Type2  Type3 Expected_Output    New
1:   Red Orange   Pink            Pink   Pink
2: Green    abc     NA             abc    abc
3:  Blue     NA     NA            Blue   Blue
4: white     NA  Green           Green  Green
5:    NA  Brown purple          purple   grey
6:    NA  black     NA           black purple
7:  grey     NA     NA            grey  black

数据

fread()用于读取示例数据集。 na.strings参数告诉fread()将"na"字符串转换为NA：

library(data.table)
DT <- fread(
"Type1   Type2   Type3       Expected_Output                 
Red     Orange  Pink        Pink                
Green   abc     na          abc             
Blue     na     na          Blue                
white    na     Green       Green               
na     Brown    purple      purple              
na     black     na         black               
grey    na       na         grey    ",
na.strings = "na")

修改

The OP has requested也应该在输出中显示完全NA的行。这可以通过更改右连接中data.table个对象的顺序来实现。在data.table语法中，X[Y]是一个右连接，它接受Y的所有行。如果需要X的所有行，则必须使用正确的联接Y[X]

library(data.table)
# new data with 8th row
DT <- fread(
  "Type1   Type2   Type3       Expected_Output                 
  Red     Orange  Pink        Pink                
  Green   abc     na          abc             
  Blue     na     na          Blue                
  white    na     Green       Green               
  na     Brown    purple      purple              
  na     black     na         black               
  grey    na       na         grey    
  na      na       na         na",
  na.strings = "na")

melt(DT[, rn := .I], id.vars = "rn", na.rm = TRUE)[
  order(variable), .(New = last(value)), by = rn][DT, on = .(rn)][, rn := NULL][]

      New Type1  Type2  Type3 Expected_Output
1:   Pink   Red Orange   Pink            Pink
2:    abc Green    abc     NA             abc
3:   Blue  Blue     NA     NA            Blue
4:  Green white     NA  Green           Green
5: purple    NA  Brown purple          purple
6:  black    NA  black     NA           black
7:   grey  grey     NA     NA            grey
8:     NA    NA     NA     NA              NA

Answer 2

您可以通过last来自dplyr内的apply来执行此操作。确保na为NA，以便na.omit忽略它们。

library(dplyr)
df[df=="na"] <- NA #change "na" to NA
df$expected2 <-apply(df[,1:3],1,function(x) last(na.omit(x)))

  Type1  Type2  Type3 Expected_Output expected2
1   Red Orange   Pink            Pink      Pink
2 Green    abc   <NA>             abc       abc
3  Blue   <NA>   <NA>            Blue      Blue
4 white   <NA>  Green           Green     Green
5  <NA>  Brown purple          purple    purple
6  <NA>  black   <NA>           black     black
7  grey   <NA>   <NA>            grey      grey

数据

df <- read.table(text="Type1   Type2   Type3 Expected_Output                 
Red     Orange  Pink        Pink                
                 Green   abc     na          abc             
                 Blue     na     na          Blue                
                 white    na     Green       Green               
                 na     Brown    purple      purple              
                 na     black     na         black               
                 grey    na       na         grey ",header=TRUE,stringsAsFactors=FALSE)

Answer 3

这是基于ifelse的另一个简单解决方案；希望对您有帮助。

create