Question

Folks, we have the following problem: we’ve got several objects containing table data that look something like this:

{'field1':'value1','field2':'value2', ...}

At some point during runtime we need to “select” data from these objects (tables) in the same way one were to query a DB (i.e. get records in this set that matches field1 == some_value and field2==some_other_value). We don’t have access to the original RDBMS or db views. We fiddled with the idea of using an intermediary DB (like sqlite) and then query it for the data as needed.

But it felt “smelly” to add another moving part to the app just for dataset querying purposes. So, my question is: is there a pythonic way to approach this? Should we bite the bullet and push the data to a DB, query the DB, then delete? Thanks in advance for your opinion and input.

The data is a list of dictionaries.

Answer 1

def one_of(alternatives):
    return lambda val: val in alternatives

def within(frm, to):
    return lambda val: frm <= val <= to

def gte(const):
    return lambda val: val >= const

def lte(const):
    return lambda val: val <= const

def exact(const):
    return lambda val: val == const


def select(data, **kwargs):
    for item in data:
        if all(chk(item[key]) for key, chk in kwargs.items()):
            yield item

data = [
    {'id': 1, 'name': 'Ann', 'age': 25},
    {'id': 2, 'name': 'Tom', 'age': 10},
    {'id': 3, 'name': 'John', 'age': 40},
    {'id': 4, 'name': 'Tom', 'age': 18},
]

for person in select(data, age=gte(15), id=within(3, 4), name=exact('Tom')):
    print person['name']