> It's more explicit than MySQL, so in that way it's better.
It sounds like you're under the impression that MySQL just makes everything case-insensitive and is silent about this? That's decidedly not the case.
MySQL 8 ships with 41 different character sets, supporting a total of 286 different collations. Collation names explicitly include "ci" (case-insensitive) vs "cs" (case-sensitive), as well as "ai" (accent-insensitive) vs "as" (accent-sensitive), and also the language/region conventions used for sorting purposes.
You can choose collation at the column-level granularity, as well as setting defaults at the table, schema, and server levels. It's completely explicit and very configurable.
Eh, just from a SQL query alone, there's no way to see that (lower(email)) is indexed, or see column data types etc. That metadata lives in the table definition, which is a normal place for it, so it seems reasonable for the collation / case-insensitivity to not be explicit in the query text.
Besides, MySQL also supports functional indexes, so you could do the (lower(email)) approach in MySQL too if you really want!
No, you can't see everything, but you can see the exact comparison it's making. Is that useful? You can decide that for yourself but I like being able to see as much of the logic (and thus verify the correctness) in the query itself. Also helps with copy/paste and some other things.
I never said you can't use functional indexes in MySQL. Someone said "you can't do this in PostgreSQL" and I just commented "here's (one way) to do this, maybe that's helpful some day". That's it.
It sounds like you're under the impression that MySQL just makes everything case-insensitive and is silent about this? That's decidedly not the case.
MySQL 8 ships with 41 different character sets, supporting a total of 286 different collations. Collation names explicitly include "ci" (case-insensitive) vs "cs" (case-sensitive), as well as "ai" (accent-insensitive) vs "as" (accent-sensitive), and also the language/region conventions used for sorting purposes.
You can choose collation at the column-level granularity, as well as setting defaults at the table, schema, and server levels. It's completely explicit and very configurable.