Hi Dan: Can you create a small example set of BLAST output, and a small version of the test code (just pull out the troublesome stuff) and share it? Barring that, if it is not possible, try using Data::Dumper on the resulting parsed document. Could you turn on the ErrorContext expat option (in XML::Parser)? Reading the rest of the man page, if you are using it in streaming mode, look carefully at the Stream_Delimiter option. Joe On Fri, 2003-08-01 at 13:33, Dan Bolser wrote: > I am now deeply frustrated. > > My XML::Parser is working almost perfectly, apart from > dumping the occasional Q instead of the correct > Hit_accession, and then somtimes 'loosing' the last > digit of the Hsp_num... > > I only notice these when they violate my primary > keys, so I don't know how many 'silent' mutations > are being introduced. > > If I put some new lines in the file, after the <!DOCYTPE, > and before <BlastOutput, other characters go missing, > but the previous ones are returned. There is a specific > number of lines which gives no detected mutations, > > but puting the file back the way it was makes the original > problem come right back - in exactly the same predetermined > way.... > > Who should I kill? > > Please don't sugest using another parser... > > Which Bioperl should I use? > > I found the great looking > > BLAST::Report > > http://ccgb.umn.edu/~crow/projects/xmlblast/ > > uses dom, which is a shame. > > Thanks anyone who can return my missing characters, > and my sanity.... > > Why would XML::Parser develope this problem is such > a derministic random way? > > ?Thanks again, Dan. > > > > _______________________________________________ > Biodevelopers mailing list > Biodevelopers at bioinformatics.org > https://bioinformatics.org/mailman/listinfo/biodevelopers -- Joseph Landman, Ph.D Scalable Informatics LLC, email: landman at scalableinformatics.com web : http://scalableinformatics.com phone: +1 734 612 4615