Advanced Queries

[codello]

Query operations should support some typical queries by implementing the following schema:

• The first parameter should accept a filter document named filter. This enables advanced queries.
• Using keyword arguments one can implement simple equality filters for arbitrary parameters.
• Some Advanced queries may be implemented using SQLAlchemy style arguemnts. E.g. age__le=20 to Filter for an age field that is less than 20. This may be generalized for arbitrary operators.
• Querying subfields is only supported by using advanced filters.
• Using keyword arguments for unknowns fields should emit a warning but not raise an error.

[codello]

In order to not accidentally hog keyword argument names this proposal will be implemented with the following changes:

• All query methods get a single filter argument. Users can pass a custom dict to take advantage of the full filtering capability.
• Alternatively users can use the Query class (a subclass of dict). This class adds the ability to easily combine queries using the operators &, |, -, and ~ (the last two of which have identical behavior). Combining a query with a dict should be possible.
• A Query can be constructed using the constructor (which behaves like the usual dict constructor), factory methods (for queries spanning multiple fields like $jsonSchema) or a special q function.
• The q function (whose name is intentionally very short) can convert keyword arguments into query documents thereby covering the most frequent cases:
  • Query by ID (or using multiple IDs)
  • Query by a combination of fields (possibly using query operators). The syntax is similar to popular ORMs using age__gt to perform a $gt search on the field age. To improve future compatibility any query operators should be supported even if they do not exist in MongoDB yet. A conversion from snake_case to camelCase should be performed automatically.