Cubrid histogram - #24
Closed
soheejung-cs wants to merge 145 commits into
Closed
Conversation
:빌드 되게 만들고 나머지 TODO 및 ASSERT 처리
- 중간부 구현: SERVER_SIDE로 가기 직전의 빌드 가능하고, 실행 가능한 상태로 중간 구현
- histogram 관련 오류 수정 (db_class fetch 방법 관련)
…/cubrid into CBRD-26217-histogram # 제목: 변경 요약 (50자 이내) (refactor, hotfix, feature) # 본문: 변경 사항에 대한 자세한 설명 # - 주요 변경 사항 1 # - 주요 변경 사항 2 # 참고: 이 라인 아래의 내용은 커밋 메시지에서 제거됩니다.
class 전체의 meta 정보가 histogram에 포함되지 않도록 변경
…ld be inaccurate when a leap year was included (CUBRID#6728) http://jira.cubrid.org/browse/CBRD-26457 * Fixed an error in adddate() and subdate() regarding the last day of a leap year. * Fixed an error in handling negative numbers in the INTERVAL syntax. * Changed the processing to be performed after type casting, depending on the single or composite unit.
…clause conditions are propagated as if they were WHERE clause conditions (CUBRID#6787) Purpose In the rewriter, qo_reduce_equality_terms is applied to all terms in the WHERE condition. For CONNECT BY without joins, however, single_table_opt is enabled and moves conditions from the START WITH clause into the WHERE clause. As a result, terms from START WITH are treated like WHERE terms and are reduced accordingly, so the query can return wrong results. To fix this, reduce_equality_terms is disabled when single_table_opt is enabled and the CONNECT BY query has a START WITH clause.
Owner
Author
|
@codex 리뷰해줘 파일별로 정리해줘 |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c21c6c0db1
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Comment on lines
+2904
to
+2908
| int min_weight = (total_pages + 99) / 100; /* ensure minimum 100 pages (rounded up) */ | ||
| int max_weight = total_pages / 5000; /* limit maximum 5000 pages */ | ||
|
|
||
| /* select the smaller value between base_weight and min_weight, and greater than max_weight */ | ||
| hsidp->sampling.weight = MAX (MIN (base_weight, min_weight), MAX (max_weight, 1)); |
There was a problem hiding this comment.
The new sampling logic uses min_weight = (total_pages + 99) / 100 and then clamps weight with MIN(base_weight, min_weight), which makes the sample size roughly total_pages / weight. This caps the sample to ≤100 pages instead of guaranteeing ≥100 pages. For example, with 150 pages, min_weight becomes 2, so the scan samples about 75 pages—below the stated “minimum 100 pages.” This under-samples mid-sized tables (≈101–299 pages) and weakens statistics accuracy. To honor the minimum, the weight should be bounded by floor(total_pages/100) (or another clamp that ensures sample_size ≥ 100 when total_pages ≥ 100).
Useful? React with 👍 / 👎.
…#6794) http://jira.cubrid.org/browse/CBRD-26507 feature: exclude auto-increment columns from db_serial view Updated the db_serial system catalog view to filter out sequences associated with auto-increment columns. This is achieved by excluding records in _db_serial that have a non-null class_name.
…ion function (CUBRID#6805) restore next field of a SP call node in the parse tree before returning from XASL conversion function
http://jira.cubrid.org/browse/CUBRIDQA-1338 * Test: Add artifact API access test in build job - Package CUBRID build as tar.gz and store as artifact - Test CIRCLECI_TOKEN API access to verify artifact retrieval - This is a prerequisite test for GlusterFS build cache implementation * Implement artifact-based build distribution for sql/medium tests - Build job now stores CUBRID.tar.gz as artifact instead of workspace - Workspace only contains BUILD_NUM (few bytes vs 248MB) - Add download-build-from-artifact command for test jobs - test_medium and test_sql now download build from artifact API - Remove attach_workspace from attach-and-run-tests command This reduces workspace transfer from 248MB to few bytes per test job. * Add GlusterFS-based build cache for shell tests - Add download-build-to-glusterfs command to save build artifacts to shared storage - Add download-build job (parallelism: 1) to download once for all shell test pods - Update workflow: shell tests now depend on download-build instead of build directly - This reduces network transfer from 15GB (50 pods x 300MB) to 300MB (1 download) * Use CIRCLE_SHA1 for build cache to enable reuse across workflow retries - Change build cache key from CIRCLE_WORKFLOW_ID to CIRCLE_SHA1 - This allows reusing the same build artifact for the same commit - Benefits: if tests fail and workflow is retried, build download is skipped - GlusterFS path: /home/build-cache/builds/{commit-hash}/CUBRID * Stop after build logs on default trigger Halt the build job after collect-build-logs when the branch name does not include /head.
… returns_null_on_function_errors=yes (CUBRID#6803) http://jira.cubrid.org/browse/CBRD-26518 * Fixed an error where adddate() would not return NULL even when return_null_on_function_errors=yes
…BRID#6811) [CMS] add overwrite_exvol_file key to createdb for ext vol creation
http://jira.cubrid.org/browse/CBRD-26491 * feature: add debugging method of bison(csql, loaddb) * refactor: remove old version bison debugging code(related with PARSER_DEBUG) * refactor: remove old version bison debugging code(related with DBG_TRACE_LEVEL)
# 제목: 변경 요약 (50자 이내) (refactor, hotfix, feature) # 본문: 변경 사항에 대한 자세한 설명 # - 주요 변경 사항 1 # - 주요 변경 사항 2 # 참고: 이 라인 아래의 내용은 커밋 메시지에서 제거됩니다.
…UBRID#6828) depending on the success or failure of optimizedb, return status correctly
…#6684) http://jira.cubrid.org/browse/CBRD-26410 Improves performance degradation of query execution when the SQL_TRACE_EXECUTION_PLAN system parameter is set.
…not change (CUBRID#6831) http://jira.cubrid.org/browse/CBRD-26535 * Assign the const keyword to variables whose values do not change
…ved (CUBRID#6833) http://jira.cubrid.org/browse/CBRD-23700 - modified to also delete lob files of the table in the following cases : * when dropping or truncating a table included lob type * when dropping a lob column - modified to also delete unnecessary directories. - add HFID values as parent directory to express the relationship between lob files and tables.
# 제목: 변경 요약 (50자 이내) (refactor, hotfix, feature) # 본문: 변경 사항에 대한 자세한 설명 # - 주요 변경 사항 1 # - 주요 변경 사항 2 # 참고: 이 라인 아래의 내용은 커밋 메시지에서 제거됩니다.
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
10 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.
http://jira.cubrid.org/browse/CBRD-XXXX
Purpose
N/A
Implementation
N/A
Remarks
N/A
Note
Adds column histograms to improve query selectivity and planning.
src/histogram/*(builder/reader/client) to create/read compact histogram blobsANALYZE TABLE ... UPDATE HISTOGRAM ... WITH <n> BUCKETS [WITH FULLSCAN]andANALYZE TABLE ... DROP HISTOGRAM; parser, lexer, semantic checks, name resolution, and execution implemented_db_histogramand viewdb_histogram; wiring in schema install/constants and unload tools; DDL logging and stmt enums updated=and range predicates; propagates histogram/null-frequency viaPT_NAMEand class stats loadingRisk: High — touches parser/optimizer/catalog and execution paths; impacts plan selection and DDL/ANALYZE behavior.
Written by Cursor Bugbot for commit c21c6c0. Configure here.