> On Mar 26, 2019, at 10:41 AM, Warren Young <war...@etr-usa.com> wrote: > > The biggest problem with these isn’t the use of JSON per se, it’s that each > one is unindexed, so they’ll all be a full table scan.
Well, we don’t know that; the question was just about whether those are reasonable queries, not whether they can be indexed. > You should be able to create indices on expressions like “json_extract(json, > '$.bar')”, but you’ll want to test it to find out whether the space it takes > is worth the speed increase. The index takes exactly the same amount of space as it would if ‘bar’ were a regular SQL column. > Even better would be to simply choose not to store all of your data in JSON > form It depends on the use case. The nice thing about storing data as JSON is that there’s total flexibility of the schema of that data; you can change the way the data is structured without having to alter the database. (Except to add new indexes as needed for performance.) —Jens _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
