How Should the Schemas Work

[kudla]

I'm submitting a...

• Bug report
• Feature request
• Question

Please how should one deal with the schemas?

Current behavior

Currently I define schema similar to how it is described in docs

Both migration table and migration content are created in the public schema.

Behavior is the same regardless of the defined schema existence.

Expected behavior

• Defined schema is created if one doesn't exist.
• Migration table and content affect defined schema

Minimal reproduction of the problem with instructions

{
    "defaultEnv": "def",
    "def": {
      "driver": "pg",
      "host": {"ENV": "PGHOST"},
      "port": {"ENV": "PGPORT"},
      "database": {"ENV": "PGDATABASE"},
      "schema": {"ENV": "PGSCHEMA"},
      "user": {"ENV": "PGUSER"},
      "table": {"ENV": "PG_MIGRATION_TABLE"},
      "password": {"ENV": "PGPASSWORD"}
    }
}

I've tried both ENV link and exact value

What is the motivation / use case for changing the behavior?

Using data separation on schemas level

Environment

db-migrate version: 1.0.0-beta.15, 0.11.12
plugins with versions: not sure I have one
db-migrate driver with versions: pg 1.2.2

Additional information:
- Node version: 14.17.0
- Platform:  Mac

Others:

[kylecombes]

I'm also wondering how one configures what schema the migration tables are stored under. (I've got an automatic API generator and I'd like to keep the migrations tables out of where it's looking.)

[kylecombes]

It looks like this issue was reported back in 2017 (#459).

[nstylo]

Same issue for me. Will this ever be tackled?