I have some questions around how stream names and schemas are handled in the Postgres source.

1. Is it the intention for the stream name to always include the schema? 2. Which schemas should discover give results for? All of them? 3. Should we offer schema selection at the source level? 4. How does escaping work for schemas and stream names? Is the value for stream name in the catalog always unescaped and unquoted?

My current understanding: 1. is yes, 2. is all, 3. ???, 4. unescaped and unquoted

1. "Is it the intention for the stream name to always include the schema? Yes. stream name for postgres database is "${schema_name}.${table_name}". When @Davin Chia (Airbyte) does his name space project the schema name will be split out into a separate field.

1. Which schemas should discover give results for? All of them? yes

underscore not dot? I see a bunch of tests with . in the schema name

bah. you're right. dot.

1. Should we offer schema selection at the source level? we could do this in the future, but i don't think i've heard anyone ask for it. so i guess i don't really have an opinion here.

1. How does escaping work for schemas and stream names? Is the value for stream name in the catalog always unescaped and unquoted? yeah. i think that's the pattern we follow. i think our contract is that connectors need to be able to deal with any uft8 string. so if they detect characters that need escaping / quoting, they are responsible for doing that themselves.

tu thanks

Which schemas should discover give results for? All of them?

Should this be ‘Yes’ within the configured DB?