[BioC] BiomaRt error: ncol(result) == length(attributes) is not TRUE

[Steffen Durinck]

Dear Quin,

I'm surprised you sometimes get different amounts of annotation and that sometimes you get clearly incorrect results with biomaRt.  Could you give examples of this so we can further investigate this.
There is no maximum on the length of identifiers and that is why this is not in the manual.  Unless you really retrieve a lot of data like all cDNA sequences for all genes at once, you might see your query fail as the download of the data is to big, but then it should return nothing at all.

Best,
Steffen



----- Original Message -----
From: Quin Wills <wills at stats.ox.ac.uk>
Date: Friday, March 28, 2008 3:15 am
Subject: Re: [BioC] BiomaRt error: ncol(result) == length(attributes) is not TRUE
To: Wolfgang Huber <huber at ebi.ac.uk>
Cc: Steffen <SDurinck at lbl.gov>, bioconductor at stat.math.ethz.ch

> Thank you Steffen and Wolfgang
> 
> I'm fairly au fait with R, and realise how tedious loops can be, 
> thanks. 
> For my needs, because of (i) the uncertain amount of annotation 
> returned 
> for some identifiers (ii) sometimes clearly incorrect annotation 
> returned,  it's been sometimes easier to run loops.
> 
> Apologies if I wasn't clear... my last uncertainty was not lack of 
> clarity on single/vector queries, rather how large a single query 
> one 
> can send, and that I can't find that info. Knowing that I can run 
> 3000 
> queries in one go without server retributions is useful, thanks so 
> much 
> Steffen. I'm sorry if it is written somewhere - but perhaps an 
> indication of what size query goes beyond sensible in the help/manual?
> 
> Apologies for the multiple email - it was a result of me receiving 
> a 
> 'failed to send' message this side- apparently it had sent.
> 
> Quin
> 
> 
> Wolfgang Huber wrote:
> > Dear Quin,
> >
> > In general it is sufficient the send the same mail only once to this
> > list, there is no added benefit in looping over the send button, and
> > indeed might collect you bad karma from all the people who have 
> to clean
> > their mailboxes.
> >
> > The manual page of the "getGene" function says
> >
> > Usage:
> >      getGene( id, type, mart)
> >
> > Arguments:
> >       id: *vector* of gene identifiers one wants to annotate
> >
> > I am not sure how this could be more clear. R is a well-developed 
> and> powerful language and many people have benefited from reading
> > introductions such as this one (on CRAN):
> > http://www.stats.bris.ac.uk/R/doc/manuals/R-intro.html
> >
> >   
> 
> -- 
> 
> * *
> 
> * *
> 
> * *
> 
> *Quin Wills*
> *DPhil candidate*
> 
> * *
> 
> *Department of Statistics*
> 
> *University** of Oxford***
> 
> *1 South Parks Road*
> *Oxford***
> 
> *OX1 3TG
> United Kingdom*
> 
> 
> 
> *01865 285 394*
> 
>