mongo-datatables 2.0.0

Server-side processing for DataTables and Editor with MongoDB

mongo-datatables

Server-side processing for jQuery DataTables with MongoDB.

Translates DataTables Ajax requests into MongoDB aggregation pipelines, handling pagination, sorting, filtering, search, SearchPanes, SearchBuilder, and Editor (full CRUD).

Installation

pip install mongo-datatables
# or
uv add mongo-datatables

Quick Start

from pymongo import MongoClient
from mongo_datatables import DataTables, DataField

db = MongoClient("mongodb://localhost:27017/")["mydb"]

data_fields = [
    DataField('title', 'string'),
    DataField('artist', 'string'),
    DataField('year', 'number'),
    DataField('genre', 'string'),
]

# args is the DataTables Ajax request body (a dict)
result = DataTables(db, 'albums', args, data_fields).get_rows()

db is any PyMongo Database object. args is the JSON body from a DataTables Ajax POST. get_rows() returns the standard server-side response:

{
    "draw": 1,
    "recordsTotal": 1000000,
    "recordsFiltered": 4821,
    "data": [...]
}

Framework Examples

Flask

@app.route('/api/data', methods=['POST'])
def data():
    return jsonify(DataTables(db, 'albums', request.get_json(), data_fields).get_rows())

FastAPI

@app.post('/api/data')
async def data(request: Request):
    return JSONResponse(DataTables(db, 'albums', await request.json(), data_fields).get_rows())

Django

class DataView(View):
    def post(self, request):
        return JsonResponse(DataTables(db, 'albums', json.loads(request.body), data_fields).get_rows())

Litestar

@post('/api/data')
async def data(request: Request) -> dict:
    return DataTables(db, 'albums', await request.json(), data_fields).get_rows()

Quart

@app.route('/api/data', methods=['POST'])
async def data():
    return jsonify(DataTables(db, 'albums', await request.get_json(), data_fields).get_rows())

---

DataField

DataField(name, data_type, alias=None) maps a MongoDB field to a DataTables column.

DataField('title', 'string')                          # basic field
DataField('release_date', 'date')                     # enables date comparison search
DataField('track_count', 'number')                    # enables numeric comparison search
DataField('PublisherInfo.label', 'string', 'label')   # nested field with UI alias
DataField('_id', 'objectid')                          # serialized as string in response

Valid types:

Type	Search behaviour	Uses index?	Operators
keyword	Exact equality match	Yes (regular index)	—
string	Case-insensitive regex (substring)	No	—
number	Exact equality or numeric comparison	Yes (regular index)	> >= < <= =
date	Date comparison (ISO YYYY-MM-DD)	Yes (regular index)	> >= < <= =
array	Regex against array elements	No	—
objectid	Serialized as string in response	—	—
boolean, object, null	Treated as string (regex)	No	—

Use keyword for categorical/code fields (country codes, status values, tags) where exact matching is always intended and index performance matters. Use string for free-text fields where substring and partial matching is useful.

DataField('country_code', 'keyword')   # country:US  →  {"country_code": "US"}  — uses index
DataField('name',         'string')    # name:york   →  regex, finds "New York", "Yorkshire"
DataField('year',         'number')    # year:>1990  →  {"year": {"$gt": 1990}}  — uses index
DataField('released',     'date')      # released:>=2020-01-01  — uses index
DataField('_id',          'objectid')  # serialized as string in response
DataField('PublisherInfo.label', 'string', 'label')  # nested field with UI alias

The alias is the name DataTables uses for the column (columns[i][data]). Defaults to the last segment of the field path (PublisherInfo.label → label).

---

Search

Search is where this library earns its keep. The global search box supports several modes:

Text index search (fast)

When a MongoDB text index exists, global search uses $text — fast even on multi-million-row collections:

db.albums.create_index([("title", "text"), ("artist", "text"), ("genre", "text")])

Without a text index, the library falls back to per-column regex (much slower on large collections).

Phrase search

Wrap in quotes for exact phrase matching:

"Dark Side of the Moon"

Multi-word AND search

With search[smart]=true (DataTables default), each word must match at least one searchable column:

pink floyd 1973   →  all three terms must appear across the row

Colon syntax — field-specific search

Target a specific field without needing a separate input:

artist:Bowie                 →  artist contains "Bowie" (string — regex)
artist:"David Bowie"         →  exact phrase in artist field
country_code:US              →  country_code equals "US" (keyword — exact, uses index)
year:1972                    →  year equals 1972 (number — exact, uses index)
year:>1990                   →  greater than
year:>=1990 year:<2000       →  combine multiple conditions (ANDed)
release_date:>2020-01-01     →  date comparison

Column search with ranges

Per-column search supports pipe-delimited min|max for numbers and dates:

1990|2000          →  1990 ≤ year ≤ 2000
2020-01-01|2020-12-31

Regex mode

Set search[regex]=true to treat the search value as a raw MongoDB regex:

^Dark             →  starts with "Dark"
(Floyd|Bowie)     →  matches either

Case sensitivity

Case-insensitive by default. Pass search[caseInsensitive]=false for case-sensitive matching. Per-column override via columns[i][search][caseInsensitive].

---

SearchPanes

No server-side configuration needed — call get_searchpanes_options() to populate panes on page load:

@app.route('/searchpanes', methods=['POST'])
def searchpanes():
    dt = DataTables(db, 'albums', request.get_json(), data_fields)
    return jsonify(dt.get_searchpanes_options())

---

SearchBuilder

Full server-side support with nested AND/OR criteria trees. Works automatically — no extra configuration needed.

---

Sorting

Multi-column sorting, ColReorder (order[i][name] name-based ordering), and orderData column redirect are all supported.

---

Custom Filters

Scope all queries to a subset of the collection by passing extra filter criteria as keyword arguments:

DataTables(db, 'albums', args, data_fields, status='active', label='Merge Records')

---

Editor

Full CRUD support for DataTables Editor.

from mongo_datatables import Editor

@app.route('/editor', methods=['POST'])
def editor():
    data = request.get_json()
    result = Editor(
        db, 'albums', data,
        doc_id=request.args.get('id'),
        data_fields=data_fields,
    ).process()
    return jsonify(result)

Editor also handles action=search for autocomplete and tags field types:

@app.route('/editor/search', methods=['POST'])
def editor_search():
    return jsonify(Editor(db, 'albums', request.get_json(), data_fields=data_fields).search())

Optional Editor parameters:

Parameter	Description
validators	dict mapping field names to callable(value) -> str|None
hooks	pre_create, pre_edit, pre_remove callables — return falsy to cancel
options	dict or zero-arg callable for select/radio/checkbox field options
dependent_handlers	dict mapping field names to callables for dependent field Ajax
file_fields + storage_adapter	file upload support (subclass StorageAdapter)
row_class, row_data, row_attr	per-row metadata (static value or callable)

---

Performance & Indexes

For large collections, indexes are critical — the library uses aggregation pipelines on every request.

Text index

db.albums.create_index([
    ("title", "text"),
    ("artist", "text"),
    ("genre", "text"),
])

With a text index, global search runs in ~100–300ms on multi-million-row collections. Without one, regex fallback can take 5–10+ seconds.

MongoDB allows only one text index per collection, but it can cover multiple fields.

To force regex search even when a text index exists (for substring matching):

DataTables(db, 'albums', args, data_fields, use_text_index=False)

Regular indexes

Create indexes for fields used in sorting, column search, or custom filters:

db.albums.create_index("year")
db.albums.create_index("artist")
db.albums.create_index([("artist", 1), ("year", -1)])  # compound

---

Advanced

pipeline_stages — inject aggregation stages ($lookup, $addFields, $unwind) before the $match, useful for computed or joined fields.

allow_disk_use=True — pass allowDiskUse to aggregation pipelines when complex filters exceed MongoDB's 100 MB in-memory limit.

get_export_data() — returns all matching rows without pagination for CSV/Excel export.

row_id, row_class, row_data, row_attr — per-row DT_Row* metadata, accepts a static value or a callable receiving the raw document.

See the full documentation for details.

---

Development

Run tests:

python -m pytest tests/

Run with coverage:

python -m pytest --cov=mongo_datatables tests/ --cov-report=term --cov-report=html

License

Released under the MIT License.