-
Notifications
You must be signed in to change notification settings - Fork 915
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[BUG] cudf::hash_join accepts null equality parameter at probe time #9155
Comments
This issue has been labeled |
Still desired for this to be fixed. Low priority. |
This issue has been labeled |
Still relevant, low priority. |
This issue has been labeled |
Still relevant, low priority. |
Closes #9155 This PR removes the probe-time `cudf::null_equality` parameter in `cudf::hash_join` to avoid potential mismatching bugs between building and probing a hash join object. Authors: - Yunsong Wang (https://github.com/PointKernel) - Jason Lowe (https://github.com/jlowe) Approvers: - Conor Hoekstra (https://github.com/codereport) - Robert (Bobby) Evans (https://github.com/revans2) URL: #10260
When constructing a
cudf::hash_join
instance, one must specify whether nulls will compare as equal as part of building the hash table. Then later the caller can specify the null equality again when invoking the probe table join methods, however it will only perform correctly if the parameter matches the value specified when the instance was constructed.cudf::hash_join
instances should remember the setting when they were constructed and use the same setting when the probing methods are called later. This prevents bugs on the part of the caller where the parameter values can mismatch between construction and probe invocations and lead to data corruption.The text was updated successfully, but these errors were encountered: