Re: [Seqan-dev] double pass record reader

[Manuel Holtgrewe <manuel.holtgrewe@fu-berlin.de>]

On 03/09/2012 08:18 AM, Bernd Jagla wrote:
> Hi,
>
> I have one comment:
>
> the link in the web page http://trac.seqan.de/wiki/Tutorial/FileIO2 that
> should link to "Double-Pass Record Reader
> <http://docs.seqan.de/seqan/dev/SPEC_Double-_Pass+_Record+_Reader.html>:" is
> not workin...

Good catch, I fixed this.

> and I have one question:
>
> is it possible to write out to a file the things done in the first pass
> of the "double pass record reader"?
>
> I would like to write a function/program that is only working on a given
> partition. If I repeat this n times I think it would be beneficial to
> put the first pass of the reading into an external program...

Hm, you could try to use MultiSeqFile, following this tutorial here:

http://trac.seqan.de/wiki/HowTo/EfficientImportOfMillionsOfSequences

This gives you random access, albeit without storing an index into your 
reads file. Currently, storing such indices is not possible.


However: I can see how this would be useful for processing chunks of 
data. Let me see whether we can cook up something reuseable in this 
direction. There is the FAI index format from samtools, but that's only 
for FASTA files.


*m