[Biojava-l] StringIndexOutOfBoundsException while parsing blast result

[Richard Holland]

Thanks for the extra info.

2008/10/1 David Toomey <dtoomey at rcsi.ie>:
> They are on the same OS. For all my tests I have run the blast search and
> parsing on the same OS. This has mostly been windows but I have also tried
> the whole thing on Linux and I get the same problem.
> I have done some more testing and I don't think the carriage return is the
> problem.
> What I have found is that if the second line is less than 11 characters the
> error is thrown. If I add 4 spaces in front of the 'GN=ISPF' on the second
> line then it is parsed correctly, like this.
>
> 2,4-cyclodiphosphate synthase OS=Plasmodium falciparum (isolate 3D7)
>    GN=ISPF
>
> I haven't figured out why it parses correctly when it is the only entry in
> the file, even without the spaces. So maybe I am still missing something.
>
> Cheers,
>
> Dave
>
> -----Original Message-----
> From: dicknetherlands at gmail.com [mailto:dicknetherlands at gmail.com] On Behalf
> Of Richard Holland
> Sent: 30 September 2008 17:31
> To: David Toomey
> Cc: biojava-l at lists.open-bio.org
> Subject: Re: [Biojava-l] StringIndexOutOfBoundsException while parsing blast
> result
>
> Sounds like it _might_ be something to do with the carriage return
> itself. Is the blast file generated on the same OS that you're running
> your analysis on? (e.g. you might run Blast on a Linux box, but
> attempt to parse the file on a Windows box?). If the two OSes are
> different, this might point to it - as Linux won't necessarily
> understand the Windows linebreaks, or vice versa, and might
> misinterpret them. When you copy the portion of the file to a new file
> on the OS you're running the analysis on, it will substitute its own
> local linebreaks and thus mask the problem.
>
> So the first thing I'd check is to what the two OSes involved are. If
> they're different, try running your analysis program on the same OS as
> the Blast output was generated on. If that does fix it, then try
> putting your Blast files through dos2unix or something similar to
> convert the linebreaks before running your analysis program.
>
> If they're the same OS, then we still have a problem!
>
> cheers,
> Richard
>
>
>
>



-- 
Richard Holland, BSc MBCS
Finance Director, Eagle Genomics Ltd
M: +44 7500 438846 | E: holland at eaglegenomics.com
http://www.eaglegenomics.com/