Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Question: through_tables without foreign key #621

@JiiimmyyN

Description

@JiiimmyyN

Hey,

I'm trying to get tags indexed on a certain type (one to many), I've ran in to some blockers and can't get this working, it seems the reason it's not working is because of the "through table" do not contain a foreign key to the user in the example below.

I've created a minimal example that shows the issue I'm having.
The reason we do not have a foreign key is because of implementation details in a larger project, so that can sadly not be changed.

Example 1

The postgres tables:

CREATE TABLE IF NOT EXISTS users (
    id SERIAL PRIMARY KEY,
    name VARCHAR(255) NOT NULL,
    email VARCHAR(255) UNIQUE NOT NULL,
);

CREATE TABLE IF NOT EXISTS tag (
    id SERIAL PRIMARY KEY,
    name TEXT NOT NULL
);

CREATE TABLE IF NOT EXISTS user_tag (
    id SERIAL PRIMARY KEY,
    user_id INTEGER,
    tag_id INTEGER REFERENCES tag(id)
);

The pgsync schema:

[
  {
    "database": "postgres",
    "index": "users",
    "nodes": {
      "table": "users",
      "columns": ["id", "name", "email"],
      "children": [
        {
          "table": "user_tag",
          "columns": [],
          "relationship": {
            "variant": "scalar",
            "type": "one_to_many",
            "foreign_key": {
              "child": ["user_id"],
              "parent": ["id"]
            }
          },
          "children": [
            {
              "table": "tag",
              "columns": ["name"],
              "label": "tag",
              "relationship": {
                "variant": "scalar",
                "type": "one_to_one",
                "foreign_key": {
                  "child": ["tag_id"],
                  "parent": ["id"]
                }
              }
            }
          ]
        }
      ]
    }
  }
]

In this example we end up with user_tag.id in the elastic documents but we want the tag.name, like the following:

"user_tag": [
      1,
      2,
      3
    ],

Example 2

This is another approach I've tried using through_tables

[
  {
    "database": "postgres",
    "index": "users",
    "nodes": {
      "table": "users",
      "columns": ["id", "name", "email"],
      "children": [
        {
          "table": "tag",
          "columns": ["name"],
          "relationship": {
            "variant": "scalar",
            "type": "one_to_many",
            "through_tables": ["user_tag"]
          },
          "transform": {
            "rename": {
              "name": "tag"
            }
          }
        }
      ]
    }
  }
]

This one failes with a clear error about foreign keys missing, I also tried adding foreign_key manually in this case without any success.

for table, cols in self.get_foreign_keys(through, node.parent).items():
                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/usr/local/lib/python3.12/site-packages/pgsync/querybuilder.py", line 273, in get_foreign_keys
     raise ForeignKeyError(
pgsync.exc.ForeignKeyError: 'No foreign key relationship between public.user_tag and public.users'

So to my final question, am I just missing something or is this not possible? Do you have any recommended workarounds to the issue?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions