Xata Lite Feature Requests

Shared clusters use uuid-ossp for uuid generation, while this covers projects where older kinds of uuids are being used this leaves out custom ID's like nanoids or UUIDv7. Normally one would add such ID via an CREATE FUNCTION statement like https://gist.github.com/kjmph/5bd772b2c2df145aa645b837da7eca74 but those statements are forbidden on shared instances. Support for custom ID's could be added by easing the CREATE FUNCTION filter and allowing IMMUTABLE STRICT LEAKPROOF functions

Here's a request from Discord: somewhat random feature request/suggestion: it would be cool if i could sort my tables in alphabetical order in the UI (on the left hand side)

Provide filter operators (gt,lt,is etc) that work against the length of array/multiple column content, in order to select records based on the number of entries in an array/multiple field. ref. https://xata.io/docs/api-reference/db/db_branch_name/tables/table_name/query#filtering

Also a bulk upsert would be nice

Currently the id field only can contain alphanumerics and '-', '_', ':', or '~'. This character set would need to be extended to incorporate items such as email addresses as part of an id. As a workaround, content that cannot be stored using the currently available character set, can be converted to base62 ( https://en.wikipedia.org/wiki/Base62 ). Base 62 characters are: capital letters A-Z, the lower case letters a-z and the numbers 0–9. Which all fit in the current id field's character set.

It's currently not possible, could be useful.

Provide analyzer configuration options to manipulate the search query tokens. For example, a configurable stop token filter ( https://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-stop-tokenfilter.html ) can remove tokens ["why" "is" "the" "in"] so that the search text "Why is the sun in the sky" effectively produces tokens ["sun", "sky"] only.

Allow tags wrap or ability to see many tags at once ✨

In speaking with a number of users, this concept of bring your own data type has come up a few times. Some folks would like the ability to define a regex and data type to validate their data with a user defined type. This would then also be typed in the TypeScript SDK.