How to split contents in a single column into two separate columns in R?

Question

I have a column in my dataframe:

Colname
20151102
19920311
20130204
>=70
60-69
20-29

I wish to split this column into two columns like:

Col1         Col2
20151102
19920311
20130204
            >=70
            60-69
            20-29

How can I achieve this result?

Answer 1

Without the need of any package:

df[,c("Col1", "Col2")] <- ""

isnum <- suppressWarnings(!is.na(as.numeric(df$colname)))

df$Col1[isnum] <- df$colname[isnum]
df$Col2[!isnum] <- df$colname[!isnum]

df <- df[,!(names(df) %in% "colname")]

Data:

df = data.frame(colname=c("20151102","19920311","20130204",">=70","60-69","20-29"), stringsAsFactors=FALSE)

Answer 2

One possible solution, the idea is to use extract from tidyr. Note that the delimiter I choose (the dot) must not appear in your initial data.frame.

library(magrittr)
library(tidyr)

df$colname = df$colname %>% 
             grepl("[>=|-]+", .) %>% 
             ifelse(paste0(".", df$colname), paste0(df$colname, ".")) 

extract(df, colname, c("col1","col2"), "(.*)\\.(.*)")
#     col1  col2
#1  222222      
#2 1111111      
#3          >=70
#4         60-69
#5         20-29

Data:

df = data.frame(colname=c("222222","1111111",">=70","60-69","20-29"))

Answer 3

Here is a single statement solution. read.pattern captures the two field types separately in the parts of the regular expression surrounded by parentheses. (format can be omitted if the Colname column is already of class "character". Also, if it were desired to have the first column numeric then omit the colClasses argument.)

library(gsubfn)
read.pattern(text = format(DF$Colname), pattern = "(^\\d+$)|(.*)", 
                   col.names = c("Col1", "Col2"), colClasses = "character")

giving:

      col1     col2
1 20151102         
2 19920311         
3 20130204         
4          >=70    
5          60-69   
6          20-29 

Note: Here is a visualization of the regular expression used:

(^\d+$)|(.*)

Debuggex Demo