1

I have some Python code that reads in a number of data files and performs some cleaning steps on them. I've been getting a very strange error message during the read_feather function call:

Process finished with exit code 135 (interrupted by signal 7: SIGEMT)

I tried googling it, of course, but none of the scenarios I found had the slightest relevant to reading feather files. I tried adding

if __name__ == '__main__':

to the main part of the code, just after all the imports, and that fixed my problem.

Here's my question: What's going on, here? Why did simply adding that line of code fix this error message? Does feather file reading use multiprocessing, and that's why I need to protect my main code?

Adrian Keister
  • 842
  • 3
  • 15
  • 33
  • Generally, the question lacks a [mcve]. In any case, https://stackoverflow.com/questions/57786389/cannot-send-torch-tensors-larger-than-approx-64mb-through-multiprocessing-pi. It would help if you elaborated what causes you ruled out already and why, just to show an effort other than that you "tried googling it". Oh, and you probably wanted to write `__name__ == '__main__'`. – Ulrich Eckhardt May 19 '22 at 16:02
  • A MRE is rather difficult to present. The error occurs during a simple `pandas.read_feather()` call, but the file being read is on the order of 3GB. When you google the error, you see essentially one or two examples of the error, and that's it. Moreover, the error message itself is not terribly helpful in helping me know what to search. My level of ignorance is at the "I don't even know what I don't know" stage, not just the "I don't know" stage. Thanks for catching the typo. – Adrian Keister May 19 '22 at 16:23

0 Answers0