向量元素到数据框的不同列

Question

我有一个 df:

   group number id
1      A   abcd  1
2      A   abcd  2
3      A   abcd  3
4      A   efgh  4
5      A   efgh  5
6      B   abcd  1
7      B   abcd  2
8      B   abcd  3
9      B   abcd  9
10     B   ijkl 10

我想做成这样：

   group number  data1 data2 data3 data4           Length
1      A   abcd      1     2     3                      3
2      A   efgh      4     5                            2
3      B   abcd      1     2     3     9                4
4      B   ijkl      10                                 1

对不起我只能这样去df2:

   group number     data               Length
1      A   abcd  c(1,2,3)                   3
2      A   efgh  c(4,5)                     2
3      B   abcd  c(1,2,3,9)                 4
4      B   ijkl  10                         1

我的代码在这里：

library(tidyverse)

df <- data.frame (group = c(rep('A',5),rep("B",5)),
                  number = c(rep('abcd',3),rep('efgh',2),rep('abcd',4),rep('ijkl',1)),
                  id = c(1,2,3,4,5,1,2,3,9,10))

df2 <- df %>%
  group_by(group,number) %>%
  nest() %>%
  mutate(data=map(data,~unlist(.x, recursive = TRUE, use.names = FALSE))，
         Length= map(data, ~length(.x)))

请随意从 df 或 df2 开始，with(out) 任何包都可以。

Answer 1

您可以将名称 count 更改为 length（另外，我更喜欢将 'space' 更改为 NA，如果要更改它，df2[is.na(df2)]='')

选项 1

df <- data.frame (group = c(rep('A',5),rep("B",5)),
                  number = c(rep('abcd',3),rep('efgh',2),rep('abcd',4),rep('ijkl',1)),
                  id = c(1,2,3,4,5,1,2,3,9,10))

df2 <- df %>%
    group_by(group,number) %>%
    mutate(data=toString(id),count=n())

library(splitstackshape)
cSplit(df2, 3, drop = TRUE,sep=',')


   group number count data_1 data_2 data_3 data_4
1:     A   abcd     3      1      2      3     NA
2:     A   efgh     2      4      5     NA     NA
3:     B   abcd     4      1      2      3      9
4:     B   ijkl     1     10     NA     NA     NA

选项 2

library(dplyr)
library(tidyr)

df2 <- df %>%
     group_by(group,number) %>%
     summarise(data=toString(id),count=n())%>%separate_rows(data)%>% mutate(Col = paste0("data", 1:n()))%>%spread(Col, data)
df2
# A tibble: 4 x 8
# Groups:   group [2]
   group number count data1 data2 data3 data4 data5
* <fctr> <fctr> <int> <chr> <chr> <chr> <chr> <chr>
1      A   abcd     3     1     2     3  <NA>  <NA>
2      A   efgh     2  <NA>  <NA>  <NA>     4     5
3      B   abcd     4     1     2     3     9  <NA>
4      B   ijkl     1  <NA>  <NA>  <NA>  <NA>    10

Answer 2

在基地 R

temp = split(df, paste(df$group, df$number))
columns = max(sapply(temp, NROW))
do.call(rbind, lapply(temp, function(a)
    cbind(group = a$group[1],
          number = a$number[1],
          setNames(data.frame(t(a$id[1:columns])), paste0("data", 1:columns)),
          length = length(a$id))
))
#       group number data1 data2 data3 data4 length
#A abcd     A   abcd     1     2     3    NA      3
#A efgh     A   efgh     4     5    NA    NA      2
#B abcd     B   abcd     1     2     3     9      4
#B ijkl     B   ijkl    10    NA    NA    NA      1

Answer 3

我必须盲目地给你，但那应该有效或接近:

library(tidyverse)
df %>%
    group_by(group,number) %>%
    mutate(key = paste0("data",row_number()),length = n()) %>%
    ungroup %>%
    spread(key,id,"")

要使其从您的嵌套数据中工作，我认为您必须将这些向量更改为 1 行 data.frames 相同的列号和名称，然后使用 unnest，这要复杂得多！ :)

Answer 4

这是一个使用data.table

的选项

library(data.table)
dcast(setDT(df), group + number~ paste0("data", rowid(group, number)), 
 value.var = 'id', fill = 0)[, 
   length := Reduce(`+`, lapply(.SD, `>`, 0)), .SDcols = data1:data4][]
#    group number data1 data2 data3 data4 length
#1:     A   abcd     1     2     3     0      3
#2:     A   efgh     4     5     0     0      2
#3:     B   abcd     1     2     3     9      4
#4:     B   ijkl    10     0     0     0      1

Answer 5

这是的变体，它计算 Length before 从长格式重塑为宽格式，并在中使用 prefix 参数调用 rowid():

library(data.table)

data.table(df)[, Length := .N, by = .(group, number)][
  , dcast(.SD, group + number + Length ~ rowid(group, number, prefix = "data"), 
          value.var = "id")]

   group number Length data1 data2 data3 data4
1:     A   abcd      3     1     2     3    NA
2:     A   efgh      2     4     5    NA    NA
3:     B   abcd      4     1     2     3     9
4:     B   ijkl      1    10    NA    NA    NA

对于漂亮的打印，NA 值可以转换为白色 space:

data.table(df)[, Length := .N, by = .(group, number)][
  , dcast(.SD, group + number + Length ~ rowid(group, number, prefix = "data"), 
          as.character, value.var = "id", fill = "")]

   group number Length data1 data2 data3 data4
1:     A   abcd      3     1     2     3      
2:     A   efgh      2     4     5            
3:     B   abcd      4     1     2     3     9
4:     B   ijkl      1    10

向量元素到数据框的不同列

Element of vector to different columns of data frame

r

vector

dataframe

tidyverse