将数据重整为无定量数据

debugcn 发表于 Dev

用户名

我想我了解一般的重塑。但是，我有一些数据需要重整为宽格式，但是我不想显示由另一个变量索引的分数或定量数据。

相反，我想根据其值将一个变量从单个变量切换为五个变量。不应索引其他变量。我希望一个变量的值形成其他五个变量，并且这些变量的值应与变量名相同。

我提供了一个之前和之后的示例。

数据：

> dput(ansscales3)

structure(list(ATID = c(33, 33, 33, 33, 33, 33, 33, 33, 33, 33, 
33, 33, 33), AnswerTypeDesc = c("VS|S|SD|VD", "VS|S|SD|VD", "VS|S|SD|VD", 
"VS|S|SD|VD", "VS|S|SD|VD", "VS|S|SD|VD", "VS|S|SD|VD", "VS|S|SD|VD", 
"VS|S|SD|VD", "VS|S|SD|VD", "VS|S|SD|VD", "VS|S|SD|VD", "VS|S|SD|VD"
), AValue = c(4, 3, 2, 1, 4, 3, 2, 1, 2, 1, 4, 3, 4), ScaleValue = c(1, 
2, 3, 4, 1, 2, 3, 4, 3, 4, 1, 2, 1), ADesc = c("Very Satisfied", 
"Satisfied", "Somewhat Dissatisfied", "Very Dissatisfied", "Very Satisfied", 
"Satisfied", "Somewhat Dissatisfied", "Very Dissatisfied", "Somewhat Dissatisfied", 
"Very Dissatisfied", "Very Satisfied", "Satisfied", "Very Satisfied"
), AOrder = c(1, 2, 3, 4, 1, 2, 3, 4, 3, 4, 1, 2, 1), StatGroup = c("AdjN", 
"AdjN", "AdjN", "AdjN", "N", "N", "N", "N", "PctNeg", "PctNeg", 
"PctPos", "PctPos", "TopBox"), Cycles = c(11, 11, 11, 11, 11, 
11, 11, 11, 11, 11, 11, 11, 11)), .Names = c("ATID", "AnswerTypeDesc", 
"AValue", "ScaleValue", "ADesc", "AOrder", "StatGroup", "Cycles"
), row.names = c(NA, -13L), class = "data.frame")

重塑为：

> dput(atids1)

structure(list(ATID = c(33, 33, 33, 33), AnswerTypeDesc = structure(c(1L, 
1L, 1L, 1L), .Label = "VS|S|SD|VD", class = "factor"), AValue = c(4, 
3, 2, 1), ScaleValue = c(1, 2, 3, 4), ADesc = c("Very Satisfied", 
"Satisfied", "Somewhat Dissatisfied", "Very Dissatisfied"), AOrder = c(1, 
2, 3, 4), Cycles = c(11, 11, 11, 11), N = c("N", "N", "N", "N"
), AdjN = c("AdjN", "AdjN", "AdjN", "AdjN"), PctPos = c("PctPos", 
"PctPos", "", ""), PctNeg = c("", "", "PctNeg", "PctNeg"), TopBox = c("TopBox", 
"", "", "")), .Names = c("ATID", "AnswerTypeDesc", "AValue", 
"ScaleValue", "ADesc", "AOrder", "Cycles", "N", "AdjN", "PctPos", 
"PctNeg", "TopBox"), row.names = c(NA, -4L), class = "data.frame")

我敢肯定这很简单，但不幸的是我无法使用reshape方法弄清楚。

SabDeM

这是一个解决方案 reshape2

（注意：我已加载dplyr以触发%>%运算符，但这只是样式代码的个人选择）

library(reshape2)
library(dplyr)
dat1 %>% dcast(... ~ StatGroup, value.var = "StatGroup", fill = "")
  ATID AnswerTypeDesc AValue ScaleValue                 ADesc AOrder Cycles AdjN N PctNeg PctPos TopBox
1   33     VS|S|SD|VD      1          4     Very Dissatisfied      4     11 AdjN N PctNeg              
2   33     VS|S|SD|VD      2          3 Somewhat Dissatisfied      3     11 AdjN N PctNeg              
3   33     VS|S|SD|VD      3          2             Satisfied      2     11 AdjN N        PctPos       
4   33     VS|S|SD|VD      4          1        Very Satisfied      1     11 AdjN N        PctPos TopBox

另一个解决方案 tidyr

 library(tidyr)
 dat1 %>% spread(StatGroup, StatGroup, fill = "")
  ATID AnswerTypeDesc AValue ScaleValue                 ADesc AOrder Cycles AdjN N PctNeg PctPos TopBox
1   33     VS|S|SD|VD      1          4     Very Dissatisfied      4     11 AdjN N PctNeg              
2   33     VS|S|SD|VD      2          3 Somewhat Dissatisfied      3     11 AdjN N PctNeg              
3   33     VS|S|SD|VD      3          2             Satisfied      2     11 AdjN N        PctPos       
4   33     VS|S|SD|VD      4          1        Very Satisfied      1     11 AdjN N        PctPos TopBox

本文收集自互联网，转载请注明来源。

如有侵权，请联系[email protected] 删除。