fondre et remodeler les colonnes avec mot de racine de colonne similaire

Question

j'ai une trame de données comme suit

id gender group Student_Math_1 Student_Math_2 Student_Read_1 Student_Read_2 
46 M  Red 23    45    37    56 
46 M  Red 34    36    33    78 
46 M  Red 56    63    58  
62 F  Blue 59            68 
62 F  Blue     68    87    73 
38 M  Red 78    57        65 
38 M  Red     75    54 
17 F  Blue 74        56    72 
17 F  Blue 75    61        79 
17 F  Blue     74    43    81 

    df = structure(list(id = c(46, 46, 46, 62, 62, 38, 38, 17, 17, 17), 
    gender = structure(c(2L, 2L, 2L, 1L, 1L, 2L, 2L, 1L, 1L, 
    1L), .Label = c("F", "M"), class = "factor"), group = structure(c(2L, 
    2L, 2L, 1L, 1L, 2L, 2L, 1L, 1L, 1L), .Label = c("Blue", "Red" 
    ), class = "factor"), Student_Math_1 = c(23, 34, 56, 59, 
    NA, 78, NA, 74, 75, NA), Student_Math_2 = c(45, 36, 63, NA, 
    68, 57, 75, NA, 61, 74), Student_Read_1 = c(37, 33, 58, NA, 
    87, NA, 54, 56, NA, 43), Student_Read_2 = c(56, 78, NA, 68, 
    73, 65, NA, 72, 79, 81)), .Names = c("id", "gender", "group", 
"Student_Math_1", "Student_Math_2", "Student_Read_1", "Student_Read_2" 
), row.names = c(NA, -10L), class = "data.frame") 

Ce que je suis en train de faire est de remodeler cet ensemble de données telles que les Student_Math_1 et Student_Math_2 colonnes sont empilées comme seule colonne Math une dessous de l'autre et de même les Student_Read_1 et Student_Read_2 colonnes sont empilées comme une seule colonne Reading comme indiqué ci-dessous

id gender group Math Index1   Reading Index2 

46 M  Red 23 Student_Math_1 45  Student_Read_1   
46 M  Red 34 Student_Math_1 36  Student_Read_1 
46 M  Red 56 Student_Math_1 63  Student_Read_1 
62 F  Blue 59 Student_Math_1   Student_Read_1     
62 F  Blue  Student_Math_1 68  Student_Read_1  
38 M  Red 78 Student_Math_1 57  Student_Read_1  
38 M  Red  Student_Math_1 75  Student_Read_1 
17 F  Blue 74 Student_Math_1   Student_Read_1     
17 F  Blue 75 Student_Math_1 61  Student_Read_1   
17 F  Blue  Student_Math_1 74  Student_Read_1  

46 M  Red 45 Student_Math_2 56  Student_Read_2 
46 M  Red 36 Student_Math_2 78  Student_Read_2 
46 M  Red 63 Student_Math_2   Student_Read_2 
62 F  Blue  Student_Math_2 68  Student_Read_2 
62 F  Blue 68 Student_Math_2 73  Student_Read_2 
38 M  Red 57 Student_Math_2 65  Student_Read_2 
38 M  Red 75 Student_Math_2   Student_Read_2  
17 F  Blue  Student_Math_2 72  Student_Read_2 
17 F  Blue 61 Student_Math_2 79  Student_Read_2 
17 F  Blue 74 Student_Math_2 81  Student_Read_2 

Sachez seulement cela peut être atteindre d avec remodeler ou fondre et passer du format large au format long, je ne sais pas comment aller au-delà. Toute aide pour réaliser cette transformation est très appréciée.

Answer 1

En utilisant melt de reshape2 et passer à des entrées appropriées id, measure.vars:

MathDF = melt(data = DF,id=c("id","gender","group"),measure.vars = c("Student_Math_1","Student_Math_2"),value.name = "Math", 
    variable.name = "Index1") 

ReadDF = melt(data = DF,id=c("id","gender","group"),measure.vars = c("Student_Read_1","Student_Read_2"),value.name = "Read", 
    variable.name = "Index2") 


mergeDF = merge(MathDF,ReadDF,by=c("id","gender","group")) 

head(mergeDF) 
# id gender group   Index1 Math   Index2 Read 
# 1 46  M Red Student_Math_1 23 Student_Read_1 37 
# 2 46  M Red Student_Math_1 23 Student_Read_1 33 
# 3 46  M Red Student_Math_1 23 Student_Read_1 58 
# 4 46  M Red Student_Math_1 23 Student_Read_2 78 
# 5 46  M Red Student_Math_1 23 Student_Read_2 NA 
# 6 46  M Red Student_Math_1 23 Student_Read_2 56 

Answer 2

Avec le tidyverse, vous pourriez gather chaque ensemble de colonnes, puis filter sur toutes les valeurs où les indices ne correspondent pas au nombre de (en supposant que vous ne voulez pas Student_*_1 et Student_*_2 combinaisons):

library(tidyverse) 

df %>% gather(Index1, Math, contains('Math')) %>% 
    gather(Index2, Reading, contains('Read')) %>% 
    filter(parse_number(Index1) == parse_number(Index2)) 

## id gender group   Index1 Math   Index2 Reading 
## 1 46  M Red Student_Math_1 23 Student_Read_1  37 
## 2 46  M Red Student_Math_1 34 Student_Read_1  33 
## 3 46  M Red Student_Math_1 56 Student_Read_1  58 
## 4 62  F Blue Student_Math_1 59 Student_Read_1  NA 
## 5 62  F Blue Student_Math_1 NA Student_Read_1  87 
## 6 38  M Red Student_Math_1 78 Student_Read_1  NA 
## 7 38  M Red Student_Math_1 NA Student_Read_1  54 
## 8 17  F Blue Student_Math_1 74 Student_Read_1  56 
## 9 17  F Blue Student_Math_1 75 Student_Read_1  NA 
## 10 17  F Blue Student_Math_1 NA Student_Read_1  43 
## 11 46  M Red Student_Math_2 45 Student_Read_2  56 
## 12 46  M Red Student_Math_2 36 Student_Read_2  78 
## 13 46  M Red Student_Math_2 63 Student_Read_2  NA 
## 14 62  F Blue Student_Math_2 NA Student_Read_2  68 
## 15 62  F Blue Student_Math_2 68 Student_Read_2  73 
## 16 38  M Red Student_Math_2 57 Student_Read_2  65 
## 17 38  M Red Student_Math_2 75 Student_Read_2  NA 
## 18 17  F Blue Student_Math_2 NA Student_Read_2  72 
## 19 17  F Blue Student_Math_2 61 Student_Read_2  79 
## 20 17  F Blue Student_Math_2 74 Student_Read_2  81 

Answer 3

Nous pouvons utiliser melt de data.table

library(data.table) 
melt(setDT(df), measure = patterns("Math", "Read"), 
value.name = c("Math", "Read"))[, Index1 := names(df)[4:5][variable] 
      ][, Index2 := names(df)[5:6][variable]][] 

---

Ou une autre option est

pat <- c("Student_Math", "Student_Read") 
cbind(df[rep(1:nrow(df), 2), 1:3], do.call(cbind, lapply(pat, 
      function(nm) melt(df[grep(nm, names(df))]))))