Zeitpunkt Nutzer Delta Tröts TNR Titel Version maxTL So 06.10.2024 00:01:07 62.148 +4 3.766.113 60,6 Fosstodon 4.2.13 500 Sa 05.10.2024 00:00:05 62.144 +1 3.764.649 60,6 Fosstodon 4.2.13 500 Fr 04.10.2024 00:00:07 62.143 +3 3.762.219 60,5 Fosstodon 4.2.13 500 Do 03.10.2024 00:00:23 62.140 0 3.759.060 60,5 Fosstodon 4.2.13 500 Mi 02.10.2024 00:00:06 62.140 +1 3.755.928 60,4 Fosstodon 4.2.13 500 Di 01.10.2024 00:01:13 62.139 +1 3.752.620 60,4 Fosstodon 4.2.13 500 Mo 30.09.2024 00:01:12 62.138 +3 3.749.484 60,3 Fosstodon 4.2.12 500 So 29.09.2024 00:01:07 62.135 -1 3.746.995 60,3 Fosstodon 4.2.12 500 Sa 28.09.2024 00:01:07 62.136 +2 3.745.047 60,3 Fosstodon 4.2.12 500 Fr 27.09.2024 00:01:07 62.134 0 3.741.982 60,2 Fosstodon 4.2.12 500
Fedor Indutny (@indutny) · 11/2022 · Tröts: 2.513 · Folger: 1.086
So 06.10.2024 18:42
It might seem far fetched, but in my particular example the random string was just an UUIDv4 that we generated for each row. Either sorting the ids for each batch, or replacing them with time-based UUIDv7 solves the problem.
If you have multiple indexes for different columns, though, you might need to order your batches by all these columns as well!
Not only random inserts ruin performance, but they also result in quadratic insert time as the db grows.
[Öffentlich] Antw.: 0 Wtrl.: 0 Fav.: 0 · via Ivory for Mac