Repository navigation
astropy.io.ascii.read - segmentation fault when using names #3545
Description
Hi
I'm trying to run some code which worked on a previous astropy version but now won't work
I'm running astropy 1.0 on python 2.6.6
I'm trying to read a 3 line file and I get a Segmentation Fault.
This ONLY occurs it seems when I use the names keyword.
Activity
martynbristow commentedon Feb 24, 2015
If anyone has any ideas to fix it would be appreciated.
It seems a major problem :/
I can't use the autoread of the header because the header has a '#' prefix which confuses the allaignment.
I'm going to try and rollback or fudge it
astrofrog commentedon Feb 24, 2015
@martynbristow - what happens if you do simply
format='tab'instead offast_tab?cc @taldcroft @mdmueller
mdboom commentedon Feb 24, 2015
Can you provide the file (or just enough of the file) so I can try to reproduce?
martynbristow commentedon Feb 24, 2015
If I use 'tab' i get stupid headers :/ because they are misaligned, i didn't make them.
Line 1 shows i can add each header in by hand but that will be a lot of rehashing because I use this in a few scripts.
The file is http://www.astro.ljmu.ac.uk/~mab/assets/catalog.csv
Fix astropy#3545: Segfault when column header is skipped
mdboom commentedon Feb 24, 2015
@martynbristow: Can you try #3546 to see if it addresses your original issue?
taldcroft commentedon Feb 24, 2015
@martynbristow - try this:
There is a subtle difference between
format='tab'anddelimiter='\t', namely that the former does not strip leading/trailing whitespace, while the latter does. In your case you need that leading\tstripped.Fix astropy#3545: Segfault when column header is skipped
Fix astropy#3545: Segfault when column header is skipped