Closed Superharz closed 4 months ago
Does #1245 fix it?
I installed uproot from the branch of your commit and the error does not appear anymore. So I think it is solved by your commit.
Thank you for your quick fix!
Would it be possible to make a new release soon that contains this fix?
Python version:
3.11.9
Numpy version:
1.26.4
Uproot version:
5.3.9
Awkward version:
2.6.5
I am confronted with a new random crash when opening and creating arrays from a new PHYSLITE file. Random means, that the error does not always happen. Most of the time no error appears and the file is read properly but sometimes it does appear.
I have further narrowed the error down on a branch that is double nested that is mostly empty (1410 out of 270000 events are not empty, the first non empty entry is at index 682) but has the shape
270000 * var * var * int32
. When I try to only read this single branch it always crashes. When I exclude this branch it does not crash. If I read this branch together with other branches it sometimes crashes and sometimes not.In the cases it does not crash the branch in question looks like this:
The branch stores the decay products of taus (one inner list for each tau, as the process has almost no true taus the branch is empty for most events). If I only read in a different tau branch (e.g. the mass) that has one less dimension, so less nesting, I never encounter this crash. The branch looks like this:
How can I prevent this crash? The branch is mostly empty, but not completely, so it makes no sense to exclude it from reading.