Add multiple DB schema support for bal persist proposal

[daneshk]

Purpose

Many database providers allow you to organize database tables into named groups. You can use this to make the logical structure of the data model easier to understand or to avoid naming collisions between tables. These groups are known as schemas in PostgreSQL, CockroachDB, and SQL Server.

This proposal introduces support for connecting to tables in multiple database schemas in bal persist. Right now, bal persist only supports connecting to tables in the default schema of the database. However, in real-world scenarios, tables are often distributed across multiple schemas in the database. When working with such existing databases, it is essential to provide a way to explicitly represent the schema of the table in the model file.

Goals

• Allow users to connect to tables in multiple schemas in the database.
• Provide a way to represent the schema of the table explicitly in the model file.
• Maintain backward compatibility with the current bal persist API.

Approach

The current bal persist model supports the following annotations to map the model to a different table name in the database. For example, the following model will be mapped to the patients table in the database.

import ballerinax/persist.sql;

@sql:Name {
    value: "patients"
}
type Patient record {|
    readonly int id;
    string name;
    int age;
    string address;
    string phoneNumber;
    Gender gender;
	Appointment[] appointment;
|};

This proposal introduces a new annotation to map the model to a different schema in the database. For example, the following model will be mapped to the patients table in the healthcare schema in the database.

import ballerinax/persist.sql;

@sql:Schema {
    value: "healthcare"
}
@sql:Name {
    value: "patients"
}
type Patient record {|
    readonly int id;
    string name;
    int age;
    string address;
    string phoneNumber;
|};

Related issues: #1327

[ThisaruGuruge]

LGTM

[gimantha]

LGTM, I have one concern though. The schema is used to group the tables or the part of the database. Usually most of the programs are dealing with just one schema. Interacting with multiple schema is a rare case. Can we provide an alternative to globally assign a schema for all the tables being used in persist module?

[daneshk]

LGTM, I have one concern though. The schema is used to group the tables or the part of the database. Usually most of the programs are dealing with just one schema. Interacting with multiple schema is a rare case. Can we provide an alternative to globally assign a schema for all the tables being used in persist module?

@gimantha Thanks for the feedback. Shall we move this discussion to #1327, please?