datalink

datalink is a python module that lets you interact with entries of SQL data as if you were simply using dictionaries.

The module is built on top of sqlalchemy and dataset, and all SQL schemas supported in sqlalchemy are supported in datalink.

Using datalink.factory, the user creates a new class, each instance of which is linked to a single entry of data stored in an SQL database. Pushing changes to this SQL entry is as simple as modifying the instance's attributes.

Loading, saving, unique identification, and database management all take place behind the scenes, so the user doesn't need to worry about databases at all.

Installation

datalink is available on the Python Package Index.

pip install datalink

Tutorial

A full tutorial is avaiable here.

Creating datalinks

If you've ever used NameTuple from the collections module, you know how to use datalink.

Let's make a template for our data and an instance of that template, backed up to a file-based sqlite database.

Our SQL table will have the columns a, b, and c with the defaults set in my_data_fields.

import datalink
my_data_fields = {'a': 0, 'b': 'a string', 'c': []}  # default entries
MyClass = datalink.factory('MyClass', 'my_table', my_data_fields,
                           database='/tmp/my.db')

A = MyClass()    

INFO | db created at: sqlite:////tmp/my.db
INFO | created table my_table
DEBUG | Creating new database entry with id ce267917-6c2e-46d1-bacc-c52affba2d2d.

We can also instantiate data in the instance declaration. We can then use the data property to expose a dict containing a map of the column name to value to make sure it has been correctly set.

B = MyClass(a=100, b='a new string')
B.data

DEBUG | Creating new database entry with id e2bd9f8c-af49-4188-96c2-b6f4ea771188.

{'a': 100,
 'b': 'a new string',
 'c': [],
 'id': 'e2bd9f8c-af49-4188-96c2-b6f4ea771188'}

Modifying data

Now let's modify some elements and see what happens. They are accessible as properties in A as determined by the defaults we gave the factory.

A.a = 5
A.c.append(None)
print(A.a, A.c)

DEBUG | Updating existing database entry for id ce267917-6c2e-46d1-bacc-c52affba2d2d.
DEBUG | Updating existing database entry for id ce267917-6c2e-46d1-bacc-c52affba2d2d.
5 [None]

Multiple updates can be pushed to the SQL database at once with the update method.

A.update(a=1000, c=[1,2,3])
A.data

DEBUG | Updating existing database entry for id ce267917-6c2e-46d1-bacc-c52affba2d2d.

{'a': 1000,
 'b': 'a string',
 'c': [1, 2, 3],
 'id': 'ce267917-6c2e-46d1-bacc-c52affba2d2d'}

Loading data

As we can see, each new instance of MyClass receives a Unique Universal Identifier (UUID), that can be used to easily fetch rows from the database as a positional argument.

print(A.id)
id_to_load = A.id
del A
C = MyClass(id_to_load)
C.data

'ce267917-6c2e-46d1-bacc-c52affba2d2d'

DEBUG | Loaded data corresponding to ID: ce267917-6c2e-46d1-bacc-c52affba2d2d

{'a': 1000,
 'b': 'a string',
 'c': [1, 2, 3],
 'id': 'ce267917-6c2e-46d1-bacc-c52affba2d2d'}

User-specified identification

Users can also specify their own identifiers in the construction of new instances of MyClass to replace the uuid.

This can be useful in design patterns that want to store easily accessible configuration or metadata.

D = MyClass('myid', a=200, b='some other string', c=[4,5,6])

DEBUG | Creating new database entry with id myid.