update benchmark post with actual final schema

content/rnd/ulid_benchmarks/index.md

@@ -2,6 +2,7 @@
 title = "A One-Part Serialized Mystery, Part 2: The Benchmarks"
 slug = "one-part-serialized-mystery-part-2"
 date = "2023-07-15"
+updated = "2023-07-21"
 [taxonomies]
 tags = ["software", "rnd", "proclamation", "upscm", "rust", "sqlite"]
 +++
@@ -289,7 +290,6 @@ dropped the total size on disk about 25% (from 17 megabytes to 13), and the perc
 database consumed by the indexes of the `watch_quests` table went from 51% to 43% (that means the
 indexes went from being about 8.6MB to 5.6MB, 35% less than when using a composite primary key).
 
-
 ``` text
 *** Indices of table WATCH_QUESTS *********************************************
 
@@ -315,6 +315,19 @@ like that if you wanted to keep track of, you know, when things were created. Ir
 to implicit integer rowid primary keys for the `watch_quests` table, I had to make sure that there
 was a `created_at` column for it. Still a win, though!
 
+### *UPDATE (2023-07-21)!*
+
+Something I realized with the "final" schema is that you could have duplicate rows, since the only
+unique field was the `rowid`. I didn't want this. So, rather than create a `unique index on
+watch_quests (user, watch)`, I [just
+added](https://gitlab.com/nebkor/ww/-/commit/c685dc1a6b08d9ff6bafc72582acb539651a350c) a `primary
+key (user, watch)`.
+
+If that looks familiar, good eye! Doing this brings the disk usage back up to 17MB in the baseline
+benchmark, but the insert rate is still the same. In the grand scheme of things, this is still not a
+lot of data, so I'll take it anyway.
+
+
 ## Next steps with IDs
 
 This project is supposed to be more than just a testbed for learning about databases and web

themes/apollo

@@ -1 +1 @@
-Subproject commit eb02b7d3c18a397fe5baa394b50fe2c199208dbe
+Subproject commit 54e1c70c93ad5fe261a00ddf697856c621d8fc87