[WIP] prototype for caching geometries #54476
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
|
This change is |
otan
changed the title
craig bot
pushed a commit
that referenced
this pull request
Jan 14, 2021
57817: colexec: add the disk-spilling to the hash aggregator r=yuzefovich a=yuzefovich **colexec: mechanical changes for the external hash aggregator** This commit performs several mechanical changes prompted by the follow up work on the external hash aggregator: - extract the arguments to `newSpillingQueue` into a struct - add `context.Context` as the first argument to `ExportBuffered` method - extract aggregator tests into global variables. It also fixes a couple of cosmetic issues with the memory account names for the external operators. Release note: None **colexec: add the disk-spilling to the hash aggregator** This commit introduces the external hash aggregator that uses the hash-based partitioner with the in-memory hash aggregator as the "main" strategy and the external sort + the ordered aggregator as the "fallback". This approach was benchmarked against simply using the external sort + the ordered aggregator, and on larger datasets the chosen approach is noticably faster. In order for the in-memory hash aggregator to be able to actually fallback we need to keep track of all of the input tuples since it is very hard to spill the intermediate results of computation. This required the usage of a spilling queue and enqueuing the copies of all input batches into it. The benchmarks show that the performance overhead of this is relatively small while the spilling queue doesn't have to spill to disk (on the order of 15-20% hit in micro-benchmarks), however, when the spilling queue needs to use the disk, the hit can be 2-3x in case when the hash aggregator itself doesn't have spill. One notable change is that because the ordered aggregator doesn't support filtering aggregation, we cannot support it in the external hash aggregator, and as a result the hash aggregation is currently not planned if filtering aggregation is requested. Another notable change is the addition of unwrapping datum when converting it to JSON using `AsJSON` - for some reason, the row engine was panicking on `TestAggregatorAgainstProcessor` test with `json_agg` function, yet I couldn't reproduce it outside of the unit test, still I believe this addition doesn't make things worse. Fixes: #42485. Release note (sql change): Hash aggregation can now spill to disk when it exhausts its memory limit when executed via the vectorized engine. 58288: builtins: Implement ST_GeneratePoints function r=otan a=mknycha This function generates pseudo-random points until the requested number are found within the input area. Release note (sql change): Implement geo builtin ST_GeneratePoints Resolves: #48942 Dependent on: #54476 Co-authored-by: Yahor Yuzefovich <yahor@cockroachlabs.com> Co-authored-by: Marcin Knychała <knychala.marcin@gmail.com>
otan
changed the title
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.
No description provided.