I have a dataframe with one column and n rows, like this one:

data.frame(rep(x=c("c","a","c","b","c","d"),times=c(1,4,1,4,1,4)))

Now, I want to split up this column of the dataframe that for every c a new column is created. The aim is to transform the dataframe that only has one column into this form:

CodePudding user response：

With tidyverse, we could create a new group everytime c appears in the x column, then we can pivot the data wide. Generally, duplicate names are discouraged, so I created a sequential c column names.

library(tidyverse)

results <- df %>% 
  group_by(idx = cumsum(x == "c")) %>% 
  filter(x != "c") %>% 
  mutate(rn = row_number()) %>% 
  pivot_wider(names_from = idx, values_from = x, names_prefix = "c_") %>% 
  select(-rn)

Output

  c_1   c_2   c_3  
  <chr> <chr> <chr>
1 a     b     d    
2 a     b     d    
3 a     b     d    
4 a     b     d  

However, if you really want duplicate names, then we could add on set_names:

purrr::set_names(results, "c")

  c     c     c    
  <chr> <chr> <chr>
1 a     b     d    
2 a     b     d    
3 a     b     d    
4 a     b     d    

Or in base R, we could create the grouping with cumsum, then split those groups, then bind back together with cbind. Then, we remove the first row that contains the c characters.

names(df) <- "c"
do.call(cbind, split(df, cumsum(df$c == "c")))[-1,]

#  c c c
#2 a b d
#3 a b d
#4 a b d
#5 a b d

CodePudding user response：

You your columns have the same number of values, as in the example given do:

unstack(df, x ~ cumsum(x=="c"))

  X1 X2 X3
1  c  c  c
2  a  b  d
3  a  b  d
4  a  b  d
5  a  b  d

You can then remove the first row