[GH-ISSUE #190] .sort causes each item to be read again #91

New issue

Closed

opened 2026-05-23 08:30:14 -06:00 by gitea-mirror · 3 comments

gitea-mirror commented

2026-05-23 08:30:14 -06:00

Owner

Originally created by @meaningg on GitHub (Dec 17, 2022).
Original GitHub issue: https://github.com/appy-one/acebase/issues/190

.sort causes each item to be read from the database a second time - is it okay ?

with .sort

without .sort

Originally created by @meaningg on GitHub (Dec 17, 2022). Original GitHub issue: https://github.com/appy-one/acebase/issues/190 .sort causes each item to be read from the database a second time - is it okay ? with .sort ![image](https://user-images.githubusercontent.com/61989272/208239938-b339f4c7-d7bc-4a00-a02d-f8bff8b21edd.png) without .sort ![image](https://user-images.githubusercontent.com/61989272/208239951-0a2c8223-65c3-4d43-adf6-3c563e071f74.png)

gitea-mirror closed this issue

2026-05-23 08:30:15 -06:00

gitea-mirror commented

2026-05-23 08:30:17 -06:00

Author

Owner

@appy-one commented on GitHub (Dec 19, 2022):

It's a bit hard to say without the rest of your code, but if you are using sort in combination with take and skip, it will first need to determine which nodes should be loaded for the result set. If you have no index on the target sort column(s), it needs to perform a "table scan" to read all sort column(s) values. Then, it will read all requested data from nodes that are in the take / skip window: those records will have effectively been read twice.

Without sort, using take / skip is a as easy as just skipping n child nodes, then reading m nodes once

@appy-one commented on GitHub (Dec 19, 2022): It's a bit hard to say without the rest of your code, but if you are using `sort` in combination with `take` and `skip`, it will first need to determine which nodes should be loaded for the result set. If you have no index on the target sort column(s), it needs to perform a "table scan" to read all sort column(s) values. Then, it will read all requested data from nodes that are in the `take` / `skip` window: those records will have effectively been read twice. Without `sort`, using `take` / `skip` is a as easy as just skipping _n_ child nodes, then reading _m_ nodes once

gitea-mirror commented

2026-05-23 08:30:18 -06:00

Author

Owner

@meaningg commented on GitHub (Dec 19, 2022):

Yes, I use a combination of .sort and .take. Now that explains it all. Thank you!

@meaningg commented on GitHub (Dec 19, 2022): Yes, I use a combination of .sort and .take. Now that explains it all. Thank you!

gitea-mirror commented

2026-05-23 08:30:19 -06:00

Author

Owner

@appy-one commented on GitHub (Dec 19, 2022):

You're welcome!

@appy-one commented on GitHub (Dec 19, 2022): You're welcome! [![Spread the word](https://user-images.githubusercontent.com/26569719/169265089-3d593555-e1ad-4390-986b-877ac2c38a47.svg)](https://twitter.com/intent/tweet?button=&url=https://github.com/appy-one/acebase&text=I'm+using+@AcebaseRealtime+in+my+project+to+make+my+life+easier!&button=) [![contribute](https://user-images.githubusercontent.com/26569719/169265318-30c4c6a5-7c89-46a0-a7a2-ef433a8192f4.svg)](https://github.com/appy-one/acebase#contributing) [![Sponsor AceBase](https://user-images.githubusercontent.com/26569719/168233053-8e56b243-4140-40ab-9a30-4cb3cc149bfe.svg)](https://github.com/sponsors/appy-one)