Documentation improvements and/or tutorial ideas

[vmihailenco]

I have some spare time to work on improving docs and/or writing tutorials for Bun/PostgreSQL/MySQL. So please leave your feedback regarding docs here or suggest ideas for tutorials or share links to good tutorials.

[frederikhors]

I would like to take advantage of the availability of a professional like you to have in the docs insights/info/tutorials/ideas on:

1. performance best practices (I know that this is a big issue and that you are already addressing it in some pages of the docs, but perhaps a dedicated section would be more appropriate and more "salable")
2. creation of foreign keys (perhaps by automating the process with a Go script to always be launched after each code base change)
3. creation of indexes suggested during development by Bun based on the use of queries

In short, for a database neophyte like me - who does not have a DBA or a department dedicated to this - it would be very useful to have these info, especially if they comes from a genius like you.

NOTE: I'm using Bun exclusively with Postgresql. But I think this applies on others DB too.

[vmihailenco]

Thanks for the feedback.

performance best practices

We have https://bun.uptrace.dev/guide/running-bun-in-production.html#database-sql and that is all I know about Bun / database/sql. I probably will write more articles about Go/Postgres performance, but it is pretty broad topic...

creation of foreign keys (perhaps by automating the process with a Go script to always be launched after each code base change) creation of indexes suggested during development by Bun based on the use of queries

This sounds more like separate tools / feature requests. And they don't make my top 10 wanted features list ;)

[frederikhors]

I should create soon an attachments system in a small application using Postgres. The attachments will be small (max 25MB).

A page explaining the pros and cons would be interesting, like this: https://www.cybertec-postgresql.com/en/binary-data-performance-in-postgresql/

[hhuseyinpay]

It would be great if you add a section about JSONB. I'm using JSONB's for custom types(particularly structs), a lot. I discovered that there is an auto marshal/unmarshal feature. In GORM it's painful, you need to implement some interfaces for every type.

[vmihailenco]

See https://bun.uptrace.dev/postgres/data-types.html#jsonb . I have plans to write an article about model inheritance and/or polymorphic models that use JSON to store key/value data. Perhaps I will mention it there.

[hhuseyinpay]

I saw that but there is no mention of structs. So I thought that there is no support for structs. Then I tried and it worked. JSON column type is not postgres specific. I don't know If there is MySQL JSON or other support in the bun, but you could mention in Extending Bun with custom types

[abfisher0417]

Has there been any conversations about adding support for being able to query additional analytic stores like Apache Druid and BigQuery? I really like the way one can elegantly write queries using Bun and think it'd be a natural fit for additional stores, assuming one writes a dialect and driver.

[vmihailenco]

@abfisher0417 that is very unlikely, but we have a separate ClickHouse client at https://github.com/uptrace/go-clickhouse. I plan to finish the docs and promote it more once Go 1.18 is released.

[GabeAboy]

@vmihailenco I haven't been able to find any information on the pgdriver's compatibility with postgres versions.

[vmihailenco]

@GabeAboy PostgreSQL protocol is very stable so pgdriver should work with any PostgreSQL version released in the past 10 years. That does not mean it supports all PostgreSQL features though.

[victorKochkarev]

It would be a good idea to add documentation about index management. I was unable to find any information about it in the documentation. The only place I found information about it was in the code example section.