用不同数量的逗号分隔多列中的行

Question

I use R and I have a dataframe with 3 columns that contains values separed with ",".

Here's what it looks like:

col_A	col_B	col_C
first_name,last_name,age	John,Appleseed,23	Steve,Jobs, 33

I want each value separed by a comma to create a new row for this value. So it should look like this:

col_A	col_B	col_C
first_name	John	Steve
last_name	Appleseed	Jobs
age	23	33

I have succeeded to perform it by doing like this:

col_A<- strsplit(df$col_A, split = ",")
col_B<- strsplit(df$col_B, split = ",")
col_C<- strsplit(df$col_C, split = ",")

df2<-data.frame(col_A= unlist(col_A),
                 col_B=unlist(col_B),
                col_C=unlist(col_C))

the problem is the table is messy: sometimes I have different number of commas, so when I use str split, I don't have the same number of elements in my lists and the data.frame() function will not work if there isn't the same number of elements. To illustrate, sometimes I will have 3 elements separed by a comma in col_A, while there are 4 commas in col_B and col_C. And vice versa. Here's an example:

col_A	col_B	col_C
first_name,last_name,age	John,Appleseed,23,	Steve,Jobs, 33,

How can I do to get rid of this problem of formatting? Adding commas before using str_split don't seem like a good solution to me.

Answer 1

You can use str_remove() across al columns to get rid of the ending commas. Then you can separate_rows() to get what you want. This will not affect the output in rows without ending commas.

library(tidyverse)

df1 <- tibble::tribble(
                        ~col_A,              ~col_B,           ~col_C,
    "first_name,last_name,age", "John,Appleseed,23", "Steve,Jobs, 33"
    )

df2 <- tibble::tribble(
                        ~col_A,               ~col_B,            ~col_C,
    "first_name,last_name,age", "John,Appleseed,23,", "Steve,Jobs, 33,"
    )

df1 %>% 
    mutate(across(.fns = ~str_remove(.x, ",$"))) %>% 
    separate_rows(everything(), sep = ",")
#> # A tibble: 3 x 3
#>   col_A      col_B     col_C  
#>   <chr>      <chr>     <chr>  
#> 1 first_name John      "Steve"
#> 2 last_name  Appleseed "Jobs" 
#> 3 age        23        " 33"

df2 %>% 
    mutate(across(.fns = ~str_remove(.x, ",$"))) %>% 
    separate_rows(everything(), sep = ",")
#> # A tibble: 3 x 3
#>   col_A      col_B     col_C  
#>   <chr>      <chr>     <chr>  
#> 1 first_name John      "Steve"
#> 2 last_name  Appleseed "Jobs" 
#> 3 age        23        " 33"

^{Created on 2021-03-02 by the reprex package (v0.3.0)}

Answer 2

Maybe you can use regmatches like below

list2DF(lapply(df, function(x) unlist(regmatches(x, gregexpr("\\w+", x)))))

which gives

       col_A     col_B col_C
1 first_name      John Steve
2  last_name Appleseed  Jobs
3        age        23    33

Data

> dput(df)
structure(list(col_A = "first_name,last_name,age", col_B = "John,Appleseed,23,,,,",
    col_C = "Steve,Jobs, 33"), row.names = c(NA, -1L), class = "data.frame")

> df
                     col_A                 col_B          col_C
1 first_name,last_name,age John,Appleseed,23,,,, Steve,Jobs, 33