Mini256
commented
1 year ago Execution Plan:
+----------------------------------+---------+-----------+-------------------------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| id | estRows | task | access object | operator info |
+----------------------------------+---------+-----------+-------------------------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Projection_14 | 200.00 | root | | gharchive_dev.github_repos.repo_name, Column#61 |
| └─Limit_17 | 200.00 | root | | offset:0, count:200 |
| └─Selection_19 | 200.00 | root | | gt(Column#61, 5000) |
| └─StreamAgg_28 | 200.00 | root | | group by:gharchive_dev.github_repos.repo_name, funcs:count(1)->Column#61, funcs:firstrow(gharchive_dev.github_repos.repo_name)->gharchive_dev.github_repos.repo_name |
| └─IndexJoin_112 | 200.00 | root | | inner join, inner:IndexReader_111, outer key:gharchive_dev.github_repos.repo_id, inner key:gharchive_dev.github_events.repo_id, equal cond:eq(gharchive_dev.github_repos.repo_id, gharchive_dev.github_events.repo_id) |
| ├─IndexReader_116(Build) | 2304.20 | root | | index:IndexFullScan_115 |
| │ └─IndexFullScan_115 | 2304.20 | cop[tikv] | table:gr, index:index_gr_on_repo_name(repo_name) | keep order:true |
| └─IndexReader_111(Probe) | 200.00 | root | partition:pull_request_event | index:IndexRangeScan_110 |
| └─IndexRangeScan_110 | 200.00 | cop[tikv] | table:ge, index:index_github_events_on_repo_id_type_action_month_actor_login(repo_id, type, action, event_month, actor_login) | range: decided by [eq(gharchive_dev.github_events.repo_id, gharchive_dev.github_repos.repo_id) eq(gharchive_dev.github_events.type, PullRequestEvent) eq(gharchive_dev.github_events.action, opened)], keep order:false |
+----------------------------------+---------+-----------+-------------------------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+May be it should use tiflash to execute it, which is only cost 15.435s:
SELECT /*+ read_from_storage(tiflash[ge, gr]) */
`gr`.`repo_name`,
COUNT(*) AS `pr_count`
FROM
`github_events` AS `ge`
INNER JOIN `github_repos` AS `gr` ON `ge`.`repo_id` = `gr`.`repo_id`
WHERE
`ge`.`type` = 'PullRequestEvent'
AND `ge`.`action` = 'opened'
GROUP BY
`gr`.`repo_name`
HAVING
`pr_count` > 5000
LIMIT
200
Hi, it's failed to execute question
Question title
The number of projects have got over 5000 PRs
Error message
Query execution timeout.
Generated SQL
Chart info