diff options
| author | Martin Kagamino Lehoux <martin@lehoux.net> | 2025-09-29 22:05:18 +0200 |
|---|---|---|
| committer | Martin Kagamino Lehoux <martin@lehoux.net> | 2025-10-01 07:51:11 +0200 |
| commit | 3d83f6e14f3524531dbc5afadb42f2946161a2f6 (patch) | |
| tree | 748d870501c37b04132f9d556ffc0263b77b0077 /content/posts | |
| parent | 4c41e477e13d37e187c8de4225c7e8056bd9ed28 (diff) | |
Update global styles
Diffstat (limited to 'content/posts')
| -rw-r--r-- | content/posts/postgres-aggregation-speed/index.md | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/content/posts/postgres-aggregation-speed/index.md b/content/posts/postgres-aggregation-speed/index.md index 29756e8..007959f 100644 --- a/content/posts/postgres-aggregation-speed/index.md +++ b/content/posts/postgres-aggregation-speed/index.md @@ -155,7 +155,7 @@ I run these five queries with `EXPLAIN (ANALYZE, COSTS, VERBOSE, BUFFERS)`, and - `HashAggregate` are different because of the choice of data processing - `composite` is the fastest - `json` is faster than `jsonb` -- `array` is faster than `object` +- `array` is faster than `object` I thought that JSONB was faster than JSON, but I now understand it is faster for things like filtering data. But in this case, I pay for encoding in the first place, and there is no gain to expect. Only when the source column is JSONB encoded, there is no cost for encoding: it is paid on write. @@ -246,5 +246,5 @@ I don't really know why `composite text` is that slow in both benchmarks. ## TODO -- [ ] Rename folder -- [ ] Resources: postgres json, postgres aggregation, psycopg
\ No newline at end of file +- [ ] Resources: postgres json, postgres aggregation, psycopg +- [ ] Charts > tables |