Filtering¶

Filtering uses the filter query parameter and is translated by the data layer.

Full JSON filter format (single condition)¶

GET /customers?filter=[{"name":"name","op":"eq","val":"John"}]

import json
from urllib.parse import quote

import httpx

flt = json.dumps([{"name": "name", "op": "eq", "val": "John"}])
url = f"http://localhost:8000/api/customers?filter={quote(flt)}"

response = httpx.get(url)
print(response.status_code, response.json())

Full JSON filter format (multiple conditions)¶

GET /customers?filter=[{"name":"status","op":"eq","val":"active"},{"name":"age","op":"ge","val":18}]

Simple filters¶

GET /customers?filter[name]=John
GET /customers?filter[name]=John&filter[status]=active

Simple filters are treated as eq comparisons.

Relationship-path filters¶

GET /customers?filter[computers.serial]=ABC-123
GET /computers?filter[owner.id]=1

Common operators (Django ORM mapping)¶

• eq, ne
• lt, le, gt, ge
• in, not_in
• like, ilike
• is_null

Logical combinations¶

Logical filter trees are supported via and, or, and not groups.

[
    {"name": "status", "op": "eq", "val": "active"},
    {
        "or": [
            {"name": "age", "op": "gt", "val": 30},
            {"name": "role", "op": "eq", "val": "admin"}
        ]
    },
    {
        "not": {
            "name": "deleted_at",
            "op": "is_null",
            "val": false
        }
    }
]

Top-level filter list items are still AND-combined in order.

Notes¶

• URL-encode JSON values in production clients.
• Supported operators are data-layer specific.