> On Sep 13, 2019, at 10:57 AM, Hick Gunter <[email protected]> wrote:
>
> This is faster if the number of keys in the list is small relative to the
> number of records in the table.
> If the number of keys is similar to the number of records in the table, then
> a simple full table scan may be faster.
Experimentally, the optimizer seems to choose an index search even with the
simpler query. I ran this on a test database with about 30k rows.
> explain query plan select * from kv_default where key in ('a','b','c')
3|0|0|SEARCH TABLE kv_default USING INDEX sqlite_autoindex_kv_default_1 (key=?)
—Jens
_______________________________________________
sqlite-users mailing list
[email protected]
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
