pysam Samfile gives error reading bamfile

[shashidhar22]

pysam Samfile give the following error when trying to read bamfile

Traceback (most recent call last):
File "sam_flanker.py", line 10, in
bam_file = pysam.Samfile(input_file, "rb")
File "csamfile.pyx", line 314, in pysam.csamfile.Samfile.cinit (pysam/csamfile.c:4820)
File "csamfile.pyx", line 491, in pysam.csamfile.Samfile._open (pysam/csamfile.c:6730)
ValueError: file header is empty (mode='rb') - is it SAM/BAM format?

[AndreasHeger]

Is your file SAM/BAM format?

[shashidhar22]

Its a BAM file. I am able to use it with samtools without any issue.

[AndreasHeger]

Thanks, I should have been more explicit. Do you have any other information that might be helpful?

Best wishes,
Andreas

[blaiseli]

I have the same problem with a bam file for which samtools view -H returns the following:

@HD VN:1.3  SO:coordinate

Strictly speaking, the header is not empty. I just have no reads in it, which may happen when the sam file resulted from a mapping of an empty file, and is a arguably legitimate situation in an analysis pipeline.

[AndreasHeger]

Thanks, I will check this.

[lynxoid]

Any updates on the issue? I am running into the same error w/ pysam 0.8.4.

[lynxoid]

Some hints as to the cause of this here: https://code.google.com/archive/p/pysam/issues/104

Workaround: set check_header=False and check_sq=False when calling fetch().

[AndreasHeger]

Not yet, will add a test case.

[AndreasHeger]

Thanks. I had a look and it seems to me that this is indeed a feature rather than a bug.
Setting check_sq=False when opening the file is sufficient. Opened, it will return 0 for
mapped and unmapped reads and fetch() will return an empty iterator.

I have added a test case covering this and modified the error message. On balance I would hesitate to disable the check at this stage - it is currently the primary way of detecting if the format is valid when opening a file.

Please reopen if you feel strongly about this.