Robust schema usage and checks for batch CDF queries #1509
Closed
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
jackierwzhang
force-pushed
the
dbr-oss1507
branch
from
a879564 to
f43699e
Compare
jackierwzhang
force-pushed
the
dbr-oss1507
branch
from
8f08eb3 to
f299dc2
Compare
| * | ||
| * However, if there are schema changes between analysis and execution, since we froze this | ||
| * schema, our schema incompatibility checks will kick in during the scan so we will always | ||
| * be safe - Although it is a notable caveat that user should be aware of because the CDC query |
There was a problem hiding this comment.
are users aware of this? should we update any documentation to make this clearer?
There was a problem hiding this comment.
Nothing is documented yet for this behavior, but we probably should cc @jose-torres
| * may break. | ||
| */ | ||
| private lazy val endingVersionForBatchSchema: Long = endingVersion.map { v => | ||
| // As defined in docs, if ending version is greater than the latest version, we will just use |
There was a problem hiding this comment.
which docs? method docs? public website docs?
There was a problem hiding this comment.
method doc, i could make it clearer.
Comment on lines
53
to
56
| sql( | ||
| s""" | ||
| |ALTER TABLE delta.`${dir.getCanonicalPath}` ADD COLUMN (name string) | ||
| |""".stripMargin) |
There was a problem hiding this comment.
can this be 1 line instead of 4?
scottsand-db
approved these changes
Dec 15, 2022
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Outdated suggestions cannot be applied.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.
Description
Properly handle read-incompatible schema changes when querying batch CDF (e.g.
table_changes()TVF, and SQL/DF APIs).Right now, batch CDF is almost always serving past data using the latest schema, but the latest schema may not be read-compatible with the data files.
this PR introduces the following:
changeDataFeed.defaultSchemaModeForColumnMappingTablethat can be set toendVersion,latestorlegacy. Iflegacyit would fallback to the current behavior in which either latest schema is used or a time-travel version schema is used, ifendVersionis set, we will use the end version's schema to serve the batch and iflatestis set, the latest schema will always be used. Note, this is orthogonal to 1), checks will be triggered all the time to ensure read-compatibility, even whenendVersionis used.versionOfis specified. Apparently ppl can time-travel the schema during querying batch CDF, this is probably unintentional, but since it exists, i explicitly blocked two options from being used concurrently.How was this patch tested?
New Unit tests.
Does this PR introduce any user-facing changes?
No