Timeline for How can I make this query more efficient?
Current License: CC BY-SA 3.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 14, 2014 at 20:35 | history | edited | Erwin Brandstetter |
edited tags
|
|
| Aug 14, 2014 at 20:04 | history | edited | ypercubeᵀᴹ | CC BY-SA 3.0 |
edited title
|
| Aug 14, 2014 at 19:43 | answer | added | Erwin Brandstetter | timeline score: 2 | |
| Aug 14, 2014 at 18:33 | vote | accept | kolrie | ||
| Aug 14, 2014 at 17:45 | comment | added | Erwin Brandstetter |
Essential information for optimization: Cardinalities of involved tables. How many different user_id in each table? How many distinct race_id in laps? What percentage of rows is after updated_at > '2014-06-13' in laps? Which are the variable parts of the query? Ideally also: (minimal) table definitions including indexes and constraints.
|
|
| Aug 14, 2014 at 17:37 | comment | added | Erwin Brandstetter |
user_id = 1 ... user_id = 148 Two different IDs in the same query? Typo? Or what is the meaning?
|
|
| Aug 14, 2014 at 9:53 | answer | added | ypercubeᵀᴹ | timeline score: 4 | |
| Aug 14, 2014 at 7:24 | comment | added | Mihai |
A derived table seems redundant,use SELECT race_id,MAX(updated_at) FROM laps WHERE user_id = 148 AND updated_at > '2014-06-13' GROUP BY race_id Also a compsoed index on races(account_id,id) might speed things up.Add last_updated_at column to it if it belongs to the races table.
|
|
| Aug 14, 2014 at 3:02 | comment | added | Craig Ringer |
Per explain.depesz.com/s/QjK, you're right, most of the time is a seqscan on laps. Is there a composite index on laps(user_id, updated_at)?
|
|
| Aug 14, 2014 at 2:29 | review | First posts | |||
| Aug 14, 2014 at 4:53 | |||||
| Aug 14, 2014 at 2:27 | history | asked | kolrie | CC BY-SA 3.0 |