2回答
- 新しい順
- 投票が多い順
- コメントが多い順
1
The default behavior of the %load magic is to not fail on error/failure. If you leave the --fail-on-failure flag off, it defaults to the False state.
Even in the event that the loader continues to execute and skips over the failures, it will end in a LOAD_FAILED state because of the errors encountered in the process. You can see the errors that are generated (and CSV lines skipped) by using the %load_status magic with the --details and --errors flags set.
回答済み 1ヶ月前
0
Seems this command doesn't have an option called --fail-on-failure
https://ipython.readthedocs.io/en/stable/interactive/magics.html#magic-load
This is not that
%load. This is the%loadfrom the graph-notebook project; https://pypi.org/project/graph-notebook/
I already did that, and actually I see that some errors happen because some nodes of the relationship do not exist. Concerning the flag --fail-on-failure and its value False by default, I understand that in this case, even with errors, my relationship should be created for the nodes that exist. Is it True?? Because this is not the case today for me. I got a LOAD FAILED and when I check if the relationship has been created, I can see that it has not been created at all, even for the nodes that exist in the DB, is that normal?
If the flag is not set, it should load whatever edges it can that have both nodes of the edge already in the graph. Another possibility is that there could be a parsing error with those lines in the CSV, preventing it from loading those edges. You might also want to set the
--modeflag toNEWto ensure the loader is passing over all of the CSV files that you are (re)attempting to load.