Fix improper handling of empty arrays in StringDimensionIndexer. #5012
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This bug was able to introduce data errors: if the input rows to an IncrementalIndex contained entirely empty arrays and single values, then upon persisting to disk, the empty arrays would be replaced with the lexicographically smallest single value, rather than nulls like they should have been.
gianm
added a commit
to gianm/druid
that referenced
this pull request
Oct 27, 2017
…che#5012) * Fix improper handling of empty arrays in StringDimensionIndexer. This bug was able to introduce data errors: if the input rows to an IncrementalIndex contained entirely empty arrays and single values, then upon persisting to disk, the empty arrays would be replaced with the lexicographically smallest single value, rather than nulls like they should have been. * Style fix. * Add tests for bitmap indexes too.
Member
|
@fjy please don't merge that fast, I was planning to review this PR. It's not a blocker for 0.11 because we are still going to submit more fixes for 0.11 next week |
Contributor
|
@leventov Please let us know sooner if you plan to review |
jihoonson
pushed a commit
that referenced
this pull request
Oct 28, 2017
…) (#5019) * Fix improper handling of empty arrays in StringDimensionIndexer. This bug was able to introduce data errors: if the input rows to an IncrementalIndex contained entirely empty arrays and single values, then upon persisting to disk, the empty arrays would be replaced with the lexicographically smallest single value, rather than nulls like they should have been. * Style fix. * Add tests for bitmap indexes too.
Member
gianm
added a commit
to implydata/druid-public
that referenced
this pull request
Nov 8, 2017
…che#5012) * Fix improper handling of empty arrays in StringDimensionIndexer. This bug was able to introduce data errors: if the input rows to an IncrementalIndex contained entirely empty arrays and single values, then upon persisting to disk, the empty arrays would be replaced with the lexicographically smallest single value, rather than nulls like they should have been. * Style fix. * Add tests for bitmap indexes too.
leventov
pushed a commit
to metamx/druid
that referenced
this pull request
Nov 13, 2017
…che#5012) (apache#5019) * Fix improper handling of empty arrays in StringDimensionIndexer. This bug was able to introduce data errors: if the input rows to an IncrementalIndex contained entirely empty arrays and single values, then upon persisting to disk, the empty arrays would be replaced with the lexicographically smallest single value, rather than nulls like they should have been. * Style fix. * Add tests for bitmap indexes too.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The issue was that adding an empty array didn't add null to the dictionary,
but it's important to add null to the dictionary because (a) it's used when
converting [] -> null for single value columns; (b) even in multi-value columns,
it's present in the inverted index and it should point back to any empty arrays.
If some other row added null to the dictionary (maybe it was a literal null,
missing value, or empty string) then everything worked out ok, but if not
then persisting would write out bad data.
This bug was able to introduce data errors: if the input rows to an
IncrementalIndex contained entirely empty arrays and single values, then
upon persisting to disk, the empty arrays would be replaced with the
lexicographically smallest single value, rather than nulls like they
should have been.
I'm not sure what patch introduced the bug, but I tested some earlier
releases, and 0.9.2 doesn't include it but 0.10.0 and 0.10.1 both do.
I marked this fix for 0.11.0 due to the potential for data errors.