[R] Converting chr to num

[Rui Barradas]

Hello,

Inline.

On 20/08/2018 01:08, Daniel Nordlund wrote:
> See comment inline below:
> 
> On 8/18/2018 10:06 PM, Rui Barradas wrote:
>> Hello,
>>
>> It also works with class "factor":
>>
>> df <- data.frame(variable = c("12.6%", "30.9%", "61.4%"))
>> class(df$variable)
>> #[1] "factor"
>>
>> as.numeric(gsub(pattern = "%", "", df$variable))
>> #[1] 12.6 30.9 61.4
>>
>>
>> This is because sub() and gsub() return a character vector and the 
>> instruction becomes an equivalent of what the help page ?factor 
>> documents in section Warning:
>>
>> To transform a factor f to approximately its original numeric values, 
>> as.numeric(levels(f))[f] is recommended and slightly more efficient 
>> than as.numeric(as.character(f)).
>>
>>
>> Also, I would still prefer
>>
>> as.numeric(sub(pattern = "%$","",df$variable))
>> #[1] 12.6 30.9 61.4
>>
>> The pattern is more strict and there is no need to search&replace 
>> multiple occurrences of '%'.
> 
> The pattern is more strict, and that could cause the conversion to fail 
> if the process that created the strings resulted in trailing spaces. 

That's true, and I had thought of that but it wasn't in the OP's problem 
description.
The '$' could still be used with something like "%\\s*$":

as.numeric(sub('%\\s*$', '', df$variable))
#[1] 12.6 30.9 61.4


Rui Barradas


> Without the '$' the conversion succeeds.
> 
> df <- data.frame(variable = c("12.6% ", "30.9%", "61.4%"))
> as.numeric(sub('%$', '', df$variable))
> [1]   NA 30.9 61.4
> Warning message:
> NAs introduced by coercion
> 
> 
> <<<snip>>>
> 
> 
> Dan
> 

---
This email has been checked for viruses by AVG.
https://www.avg.com