如何将矩阵的列分为两列？

Question

I have matrix, which one of the columns of it has IDs which sperated with ",". I just want to split that cloumn into two columns st each new column just have one part of IDs. what is the easiest way to do it ?

My matrix is:

> L
     a    u                    
[1,] "10" "mature,MIMAT0000062"
[2,] "20" "stemloop"           
[3,] "40" "mature,MIMAT0000062"

and the expected output is :

> k
     a    u          v             
[1,] "10" "mature"   "MIMAT0000062"
[2,] "20" "stemloop" "NA"          
[3,] "40" "mature"   "MIMAT0000062"
> 

Edit:

Now I have to split this matrix into two matrix based on the column with "NA" values, one with all "NA" and other without "NA".

Input:

>k
       a    u         v
[1,] "10" "mature"    "MIMAT0000062"
[2,] "20" "stemloop"  "NA"
[3,] "40" "mature_2"  "MIMAT0000043"

Output should be like,

>k1
       a    u         v
[1,] "10" "mature"    "MIMAT0000062"
[2,] "40" "mature_2"  "MIMAT0000043"

>k2
       a    u         v
[1,] "20" "stemloop"  "NA"

Answer 1

I have a function called cSplit that is quite fast and deals with these types of problems very easily.

Here are a few examples of the function in use, along with some different cases to consider:

Your existing sample data:

M1 <- cbind(a = c(10,20,40), 
            u = c("mature,MIMAT0000062", 
                  "stemloop", "mature,MIMAT0000062"))
cSplit(data.frame(M1), "u", ",")
#     a      u_1          u_2
# 1: 10   mature MIMAT0000062
# 2: 20 stemloop           NA
# 3: 40   mature MIMAT0000062

One "u" value with a comma at the start:

M2 <- cbind(a = c(10,20,40), 
            u = c(",MIMAT0000062", 
                  "stemloop", "mature,MIMAT0000062"))
cSplit(data.frame(M2), "u", ",")
#     a      u_1          u_2
# 1: 10          MIMAT0000062
# 2: 20 stemloop           NA
# 3: 40   mature MIMAT0000062

One "u" value that splits into 3 columns:

M3 <- cbind(a = c(10,20,40), 
            u = c("mature,MIMAT0000062", 
                  "stemloop,,something", "mature,MIMAT0000062"))
cSplit(data.frame(M3), "u", ",")
#     a      u_1          u_2       u_3
# 1: 10   mature MIMAT0000062        NA
# 2: 20 stemloop              something
# 3: 40   mature MIMAT0000062        NA

Answer 2

This works when the values are comma separated:

sep_cols = matrix(unlist(strsplit(as.character(L$u), ",")), ncol = 2)
new_L = cbind(L, sep_cols)

Answer 3

A different way..

a <- c(10,20,40)
u <- c("mature,MIMAT0000062", "stemloop", "mature,MIMAT0000062")
L <- data.frame(a,u)     #better use a data.frame


v <- strsplit(as.character(L$u), ",")
L$u <- sapply(v, `[`, 1)
L$v <- sapply(v, `[`, 2)

> L
#   a        u            v
#1 10   mature MIMAT0000062
#2 20 stemloop         <NA>
#3 40   mature MIMAT0000062

Answer 4

A two liner:

L$v =sapply(strsplit(as.character(L$u),","), "[", 2)
L$u =sapply(strsplit(as.character(L$u),","), "[", 1)
#L
#   a        u            v
#1 10   mature MIMAT0000062
#2 20 stemloop         <NA>
#3 40   mature MIMAT0000062

Answer 5

Another alternative using reshape2::colsplit as joran suggested:

library(reshape2)
k = cbind(a =L$a,colsplit(L$u,",",c("u","v")))
#k
#   a   u       v
#1  10  mature  MIMAT0000062
#2  20  stemloop     
#3  40  mature  MIMAT0000062