Adding `sort' in R makes the output fail

Adding `sort' in R makes the output fail

[Sebastien Vauban]

Hello,

The following code does return a corrupted answer, while it's supposed
to return a sorted dataframe.

--8<---------------cut here---------------start------------->8---
#+BEGIN_SRC R :results value replace
con <- textConnection("
index	email
A	abc-hcDgGtZH8xNBDgjK7y7TUQ@public.gmane.org
B	def-hcDgGtZH8xNBDgjK7y7TUQ@public.gmane.org
C	def-hcDgGtZH8xNBDgjK7y7TUQ@public.gmane.org
D	ghi-hcDgGtZH8xNBDgjK7y7TUQ@public.gmane.org
E	abc-hcDgGtZH8xNBDgjK7y7TUQ@public.gmane.org
F	def-hcDgGtZH8xNBDgjK7y7TUQ@public.gmane.org")
df <- read.table(con, header = TRUE)
sort(table(df$email))
#+END_SRC

#+results:
| 1 |
| 2 |
| 3 |
--8<---------------cut here---------------end--------------->8---

See video at http://screencast.com/t/4u4Sj1Mkycwp to see the effect of
adding `sort', and see the output in RStudio (correct in both cases:
with and without `sort') on the right side of the video.

What's wrong here?

Best regards,
  Seb

-- 
Sebastien Vauban

Re: Adding `sort' in R makes the output fail

[Sebastien Vauban]

"Charles C. Berry" wrote:
> On Thu, 23 Oct 2014, Sebastien Vauban wrote:
>>
>> The following code does return a corrupted answer, while it's supposed
>> to return a sorted dataframe.
>
> Not so. See below.

So, that's RStudio that's wrong, in some way?

>> --8<---------------cut here---------------start------------->8---
>> #+BEGIN_SRC R :results value replace
>> con <- textConnection("
>> index	email
>> A	abc-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
>> B	def-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
>> C	def-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
>> D	ghi-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
>> E	abc-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
>> F	def-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org")
>> df <- read.table(con, header = TRUE)
>> sort(table(df$email))
>> #+END_SRC
>>
>> #+results:
>> | 1 |
>> | 2 |
>> | 3 |
>> --8<---------------cut here---------------end--------------->8---
>
> You created a `table' object, then you sorted it creating an `array' 
> object- an object of a different class.
>
> Is this what you want?
>
> #+BEGIN_SRC R :results value replace
> con <- textConnection("
> index	email
> A	abc-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
> B	def-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
> C	def-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
> D	ghi-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
> E	abc-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
> F	def-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org")
> df <- read.table(con, header = TRUE)
> as.table(sort(table(df$email)))
> #+END_SRC
>
> #+RESULTS:
> | ghi-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org | 1 |
> | abc-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org | 2 |
> | def-hcDgGtZH8xNBDgjK7y7TUQ-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org | 3 |

Yes, it is. Thanks!

Best regards,
  Seb

-- 
Sebastien Vauban

Re: Adding `sort' in R makes the output fail

[Charles C. Berry]

On Thu, 23 Oct 2014, Sebastien Vauban wrote:

> "Charles C. Berry" wrote:
>> On Thu, 23 Oct 2014, Sebastien Vauban wrote:
>>>
>>> The following code does return a corrupted answer, while it's supposed
>>> to return a sorted dataframe.
>>
>> Not so. See below.
>
> So, that's RStudio that's wrong, in some way?
>

Hmmm. Not exactly wrong.

There are a few wrinkles:

org-babel-R-write-object-command specifies a call to 
'write.table(object)', where `object' is what is to be printed.

`write.table' coerces whatever it is trying to print to a 
`data.frame'. See ?write.table

There is a method for `table' in `as.data.frame'. Try

  as.data.frame(table(1:3))

in the console or R gui to get a sense of what happens.

There is also an `array' method for `as.data.frame', but with a one 
dimensional array it passes control to the `vector' method. Try:

as.data.frame(sort(table(1:3)))

So it is really differences in the result of as.data.frame() that you were 
seeing.

I DK what RStudio does, but it must try harder a little harder to present 
1 dimensional arrays.

HTH,

Chuck