Question

我有一个宽格式的数据框

structure(list(Id = structure(2:1, .Label = c("Jake", "Jim"
), class = "factor"), Q2.2 = c(NA, NA), Q2.4 = c(NA, NA), Q2.5 = c(NA, 
NA), Q2.6 = c(NA, 4L), Q3.2 = c(NA, NA), Q3.4 = c(NA, NA), Q3.5 = c(NA, 
NA), Q3.6 = c(4L, NA), Q4.2 = c(NA, NA), Q4.4 = c(2L, NA), Q4.5 = c(NA, 
3L), Q4.6 = c(4L, NA), Q5.2 = c(NA, NA), Q5.4 = c(2L, NA), Q5.5 = c(NA, 
NA), Q5.6 = c(NA, NA)), class = "data.frame", row.names = c(NA, 
-2L))

我试图仅根据句点后的数字将其转换为长格式的数据框。因此，带有.2的列中的所有值都应放在一个列中，而带有.4的列中的所有值都应放在另一列中。每个后缀均如此。为了更清楚地解释，我想要看起来像这样的

Answer 1

我认为OP的问题在评论中张贴的重塑问题上有更多的细节

您可以使用data.table：

library(data.table)                                                                                                                                                                                                                                                                                                                                       

setDT(dt) # Convert to data.table object

# Melt it into long format
dt <- melt(dt, id.vars = "Id")
dt[, variable := as.character(variable)]
dt[, value := as.character(value) ]

# make a substring of the type ID
type_id <- sapply((strsplit(dt$variable,
 split = ".", fixed = TRUE)),
 '[[', 2) # This is a subsetting function passed to sapply() - gets the second element of the output list

type_id <- paste0("type", type_id)

# Make a substring of the Quarter ID
quarter_id <- sapply((strsplit(dt$variable, split = ".", fixed = TRUE)), '[[', 1)

# Use data.table's dcast() to reshape the data again to get the shape you require
dt[, type_id := type_id]
dt[, quarter_id := quarter_id]
dcast(dt, Id + quarter_id ~ type_id, value.var = "value")

## Output
    Id quarter_id type2 type4 type5 type6
1: Jake         Q2  <NA>  <NA>  <NA>     4
2: Jake         Q3  <NA>  <NA>  <NA>  <NA>
3: Jake         Q4  <NA>  <NA>     3  <NA>
4: Jake         Q5  <NA>  <NA>  <NA>  <NA>
5:  Jim         Q2  <NA>  <NA>  <NA>  <NA>
6:  Jim         Q3  <NA>  <NA>  <NA>     4
7:  Jim         Q4  <NA>     2  <NA>     4
8:  Jim         Q5  <NA>     2  <NA>  <NA>

Answer 2

@rove击败了我，但想发布另一种方法，只要我拥有80％即可：

library(reshape2)

x.melted <- melt(x, id.vars=c("Id"))

x.melted$quarter.id <- gsub("\\..*","", x.melted$variable)

x.melted$number.after.period <- gsub("Q[0-9]\\.", "", x.melted$variable)

x.casted <- dcast(x.melted, Id + quarter.id ~ number.after.period, )

colnames(x.casted)[3:ncol(x.casted)] <- paste("Type.", 
                                  colnames(x.casted)[3:ncol(x.casted)], sep="")

x.casted

    Id quarter.id Type.2 Type.4 Type.5 Type.6
1 Jake         Q2     NA     NA     NA      4
2 Jake         Q3     NA     NA     NA     NA
3 Jake         Q4     NA     NA      3     NA
4 Jake         Q5     NA     NA     NA     NA
5  Jim         Q2     NA     NA     NA     NA
6  Jim         Q3     NA     NA     NA      4
7  Jim         Q4     NA      2     NA      4
8  Jim         Q5     NA      2     NA     NA

R根据列名中的数字从宽到长格式

2 个答案: