Adds support for ZSTD encoding in schema-ddl #262
Closed
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Outdated suggestions cannot be applied.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.
This PR adds support for ZSTD encoding (#237)
The first commit modifies the default for VARCHAR (and TIMESTAMP) columns to preferentially now use ZSTD over LZO for these column types. It doesn't yet override suggestions for other data types though this would likely be useful. I'm not sure if this is something we'd like to test before going down this route (e.g., how well ZSTD functions on INTs that are uniformly distributed, normally distributed etc).
The second commit modifies the encodings on the self and parent columns from RUNLENGTH to ZSTD. For columns that primarily contain one value (such as
schema_vendor
) this change makes little difference to space on disk but for other columns, such asschema_version
this can make a significant difference. In a small sample of ~10 million rows if two schemas are in use simultaneously and we make the assumption that the schema used is independent and identically distributed (over time) this column is approximated 2/3 of the size on disk using ZSTD when compared to RUNLENGTH.