If a JSON document contains strings that can't fit in a single chunk, use the naive Blob chunker instead of the smart JSON chunker. #8723
Conversation
This file contains 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
|
@nicktobey DOLT
|
…dd a new location type representing the inside of a string.
…nStringError for both keys and values in JSON.
…ocument as a non-indexed BLOB
…w that older versions can read them.
nicktobey
force-pushed
the
nicktobey/json-big
branch
from
7dd4cd0 to
82916cd
Compare
nicktobey
changed the title
…iddle of a string.
…likely a bug in handling JSON metadata. Emit a warning, then continue, treating the document as a blob and ignoring JSON metadata.
fulghum
approved these changes
Jan 9, 2025
|
@nicktobey DOLT
|
|
@nicktobey DOLT
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 JSON chunker never creates a chunk boundary inside of a string.
Originally, this PR added functionality to allow the JSON chunker to split JSON document inside a string. This was supposed to be safe and backwards compatible, because older versions of Dolt reading documents written by newer versions of Dolt are supposed to fall back on ignoring JSON document metadata if they don't understand it and treat the document like a blob.
However, tests revealed that older clients were not checking for this in enough places and would hang when trying to read documents written with this fix. This PR also contains fixes to check the JSON metadata in more places... but this doesn't do anything for existing Dolt servers running older versions.
So instead, this PR detects when a document contains strings that exceed some limit, and instead the writer falls back on writing the document as a plain blob without metadata. The limit is currently 32KB, but can be raised in the future.
I chose to keep the logic for splitting JSON documents inside a string, although the chunker doesn't currently use it, since we may decide to enable it in the future.