Queryable Dictionaries

A wrapper to query a list of python dictionaries.

Have you ever wanted to have a SQL-like interface for a list of dictionaries? No? Well here you go anyway!

I'll admit, I don't know that this is anyone's most serious problem, but I did find myself thinking that I wish I had a nicer syntax for searching through a list of dictionaries during work on these scripts.

Features

Given a list of dictionaries, say:

d_list = [
    {"id": 1, "name": "Leo", "height": 1.9, "retired": True},
    {"id": 2, "name": "Szymon", "height": 1.72, "retired": False},
    {"id": 3, "name": "Giorgi", "height": 1.77, "retired": False},
    # ...
]

You can use two different SQL-like interfaces to query the data.

Example 1:

query_dicts(d_list).select("id", "name").where("retired")
> [{"id": 1, "name": "Leo"}]

query_dicts(d_list).select("id").where("height > 1.75")
> [{"id": 1}, {"id": 3}]

query_dicts(d_list).select("name", "height").where({"id": lambda x: x%2 == 0})
> [{"name": "Szymon", "height": 1.72}]

Reading left-to-right, this is easy to understand.

• "Query dictionaries <specified list of dictionaries>."
• "Select <specified fields>, where <filters are true>."

Example 2:

select_fields("id", "name").from_dicts(d_list).where("retired")
> [{"id": 1, "name": "Leo"}]

select_fields("id").from_dicts(d_list).where("height > 1.75")
> [{"id": 1}, {"id": 3}]

select_fields(
  "name",
  "height"
).from_dicts(
  d_list
).where(
  {"id": lambda x: x < 3},
  "height < 1.8",
)
> [{"name": "Szymon", "height": 1.72}]

Reading left-to-right, this is even easier to understand.

• "select fields <specified fields>"
• "from dictionaries <specified list of dictionaries>"
• "where "<filters are true>"

I kept the parsing very simple for the filter statements. Nested operations are not supported at this time, but maybe in the future.

Implementation

Implemented in Python3, but I think it should be compatible with 2.7. This uses modules re, functools, and operator.

There is nothing further to install.

re.finditer can be a little confusing at first. Please see this documentation.

Testing

Tests were written to leverage unittest. To run all tests, use python3 -m unittest discover.

Extensions

As mentioned above, the next feature to be included is nested operations in the filter statements. Once the filter parsing logic is completed, I would like to extend this interface to Pandas DataFrames, Matrices, and other collections.

---

Thanks ✌️