Question

I'm having a dataset of 15 million rows with single column. It looks like,

x_raw
A1
A2
A3
A4
B1
B2
B3
B4
C1
C2

I want to convert it to

A1 A2 A3 A4
B1 B2 B3 B4
C1 C2 C3 C4

I was trying with 'for' loop, that will transpose every 4 rows, and add them to a 'final' dataframe, but as the dataset is too large, it'll iterate almost 2.7 million times which is not working that efficiently. Is there any other method or anything that I can use to do it efficiently?

Answer 1

Here is one option with tidyverse where the separate the 'x_raw' into two column and then spread to 'wide' format

library(dplyr)
library(tidyr)
separate(df1, x_raw, into = c('x', 'rn'), sep="(?=\\d+)", remove = FALSE) %>%
       spread(rn, x_raw) %>% 
       select(-x)
#   1  2    3    4
#1 A1 A2   A3   A4
#2 B1 B2   B3   B4
#3 C1 C2 <NA> <NA>

Or if the number of elements are always 4, then we can also do

as.data.frame(matrix(df1$x_raw, ncol =4, byrow = TRUE), stringsAsFactors=FALSE)

Answer 2

if you just want to convert to a four column data frame:

as.data.frame(matrix(df$x_raw,ncol=4,byrow = T))

Answer 3

见这个，

x_raw <- c("A1","A2","A3","A4","B1","B2","B3","B4","C1","C2","C3","C4","D1","D2","D3","D4")
x <- as.table(matrix(x_raw,ncol=4,byrow = T))
rownames(x) <- NULL
colnames(x) <- NULL
print(x)

它返回：

     [,1] [,2] [,3] [,4]
[1,] A1   A2   A3   A4
[2,] B1   B2   B3   B4  
[3,] C1   C2   C3   C4 
[4,] D1   D2   D3   D4

Answer 4

Expand the length to the next block of 4 values, and put it in a matrix:

matrix(`length<-`(dat$x_raw, (nrow(dat) %/% 4 + 1) * 4), ncol=4, byrow=TRUE)

#     [,1] [,2] [,3] [,4]
#[1,] "A1" "A2" "A3" "A4"
#[2,] "B1" "B2" "B3" "B4"
#[3,] "C1" "C2" NA   NA

Convert specific number of rows to columns in R and repeat the process for a large dataset

4 个答案: