Opening a SAM or BAM file without SQ fields in the header

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Sam file without SQ fields in the header

What is the expected output? What do you see instead?
The sam file opens without a problem, but instead I see:
    infile = pysam.Samfile(opts.filename,'rb')
  File "csamtools.pyx", line 459, in csamtools.Samfile.__cinit__ (pysam/csamtools.c:4992)
  File "csamtools.pyx", line 591, in csamtools.Samfile._open (pysam/csamtools.c:6438)
ValueError: file header is empty (mode='rb') - is it SAM/BAM format?

This is extremely useful as the output of many pipelines produce unmapped BAMs, 
and it is necessary to manipulate such data without a reference.

Original issue reported on code.google.com by nilsho...@gmail.com on 17 Oct 2012 at 4:31

[GoogleCodeExporter]

Tricky. Basically, iterating in a bam file without SQ lines works only if there 
are only unmapped reads. If there are mapped reads, a segfault occurs within 
C-samtools.

I have kept the check for header and sq lines, but added the options
check_header and check_sq to turn checking off.

Best wishes,
Andreas

Original comment by andreas....@gmail.com on 20 Nov 2012 at 7:14

• Changed state: Fixed