[R] crazy loop error.

[Bert Gunter]

Well, I'm not Prof. Ripley, but the answer is: Look at the code.
seq_len, seq.int, and seq_along call Primitives, which are implemented
in C, and therefore MUCH faster than seq(), which is implemented as
pure R code (and is also a generic, so requires method dispatch).
Though for small n (up to a few thousand, say), it probably doesn't
make much difference.(Here, to be corrected by Prof. Ripley is
needed).

-- Bert

On Tue, Jan 25, 2011 at 2:22 AM, Ivan Calandra
<ivan.calandra at uni-hamburg.de> wrote:
> Mr Ripley,
>
> May I ask why seq_len() and seq_along() are better than seq()?
>
> Thanks,
> Ivan
>
> Le 1/25/2011 09:58, Prof Brian Ripley a écrit :
>>
>> On Tue, 25 Jan 2011, Petr Savicky wrote:
>>
>>> On Mon, Jan 24, 2011 at 11:18:35PM +0100, Roy Mathew wrote:
>>>>
>>>> Thanks for the reply Erik, As you mentioned, grouping consecutive
>>>> elements
>>>> of 'a' was my idea.
>>>> I am unaware of any R'ish way to do it. It would be nice if someone in
>>>> the
>>>> community knows this.
>>>>
>>>> The error resulting in the NA was pretty easy to fix, and my loop works,
>>>> but
>>>> the results are still wrong (new script below).
>>>> Ideally it should print single "hello" for the single letters and
>>>> grouped '3
>>>> hellos' for the fives, grouped '2 hellos' for the sixes etc.
>>>>
>>>> Based on the run results, if the value of n is being tracked, it changes
>>>> quite unpredictably.
>>>> Can someone explain how the value of n changes from end of the loop to
>>>> the
>>>> top without anything being done to it?
>>>
>>> Hi.
>>>
>>> A for-loop in R is different from a for-loop in C. It is similar
>>> to foreach loop in Perl. If v is a vector, then
>>>
>>>  for (n in v)
>>>
>>> first creates the vector v and then always performs length(v) iterations.
>>> Before iteration i, n is assigned v[i] even if n is changed in the
>>> previous iteration.
>>
>> And also if v is changed during the loop.
>>
>>> If you want to control the loop variable during execution, it is possible
>>> to use a while loop, where you have full control. While loop may be
>>> better
>>> also if v has a very large length, since, for example
>>>
>>>  for (n in 1:1000000)
>>>
>>> creates a vector of length 1000000 in memory.
>>>
>>> It should also be noted that the for-loop
>>>
>>>  for (n in 1:k)
>>>
>>> performs 2 iterations, if k is 0, since 1:0 is a vector of length 2.
>>> If k may be 0, then it is better to use
>>>
>>>  for (n in seq(length=k))
>>>
>>> since seq(length=0) has length 0.
>>
>> Since you keep mentioning that, it is actually much better to use
>> seq_len(k) (and seq_along(x) instead of your earlier recommendation of
>> seq(along=x)).  And if you are using seq() in other cases in programs,
>> consider seq.int() instead.
>>
>>> Hope this helps.
>>>
>>> Petr Savicky.
>>
>
> --
> Ivan CALANDRA
> PhD Student
> University of Hamburg
> Biozentrum Grindel und Zoologisches Museum
> Abt. Säugetiere
> Martin-Luther-King-Platz 3
> D-20146 Hamburg, GERMANY
> +49(0)40 42838 6231
> ivan.calandra at uni-hamburg.de
>
> **********
> http://www.for771.uni-bonn.de
> http://webapp5.rrz.uni-hamburg.de/mammals/eng/1525_8_1.php
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>



-- 
Bert Gunter
Genentech Nonclinical Biostatistics