Document Databases

Document (aka Document-Oriented) databases store semi-structured data, which has these characteristics:

flexible data format schema:; different records in the same collection can have different data formats
flexible data format modification: the format of records can be changed without having to modify a controlling schema
queries without SQL: queries are performed based on data objects, such as JSON

Processes

Generalized Data Pipeline processes can be expressed as follows:

$\\\begin{pmatrix}h,p,n,c,o, \\ \left \{q_{f1}:q_{v1},... ,q_{fi}:q_{vi}\right \}, \\ \left \{d_{f1}:d_{v1},... ,d_{fj}:d_{vj}\right \} \end{pmatrix} \Rightarrow D\left ( h,p,n,c,o,q,d \right ) \Rightarrow \left \{r_{f1}:r_{v1},... ,r_{fk}:r_{vk}\right \} \\\\\\\mathrm{\textup{where:}}$

$\\\\h\,\,\,\,\,\,\mathrm{\textup{database host}} \\p\,\,\,\,\,\,\mathrm{\textup{database port}} \\n\,\,\,\,\,\mathrm{\textup{database name}} \\c\,\,\,\,\,\,\mathrm{\textup{database collection}} \\o\,\,\,\,\,\,\mathrm{\textup{database operation}} \\q\,\,\,\,\,\,\mathrm{\textup{query parameters}} \\q_{f}\,\,\,\,\mathrm{\textup{query field name}} \\q_{v}\,\,\,\,\mathrm{\textup{query field value}} \\i\,\,\,\,\,\,\,\mathrm{\textup{number of query parameters}} \\d\,\,\,\,\,\,\mathrm{\textup{data parameters}} \\d_{f}\,\,\,\,\mathrm{\textup{data field name}} \\d_{v}\,\,\,\,\mathrm{\textup{data field value}} \\j\,\,\,\,\,\,\,\mathrm{\textup{number of data parameters}} \\r\,\,\,\,\,\,\,\mathrm{\textup{return parameters}} \\r_{f}\,\,\,\,\mathrm{\textup{return field name}} \\r_{v}\,\,\,\,\mathrm{\textup{return field value}} \\k\,\,\,\,\,\,\mathrm{\textup{number of return parameters}} \\D\,\,\,\,\mathrm{\textup{document database functions}}$

Database Related Parameters

h - host: IP address or URL for the database
p - port: server listening port
n - name: such as ‘sales’
c - collection: such as ‘customers’, ‘products’
o - operation: action to be performed such as create, read, update, delete

Data Related Parameters

q - query parameters: identify existing database records for retrieval, update, or delete
d - data parameters: provide input data for addition or update
r - return values: provide key, value pairs of return values from operations such as retrieval
f - field name: such as ‘street’
v - value: such as ‘126 Main’
i - number of query parameters: enough key/value pairs to find the desired database documents
j - number of data parameters: provide key/value pairs for data to be added, inserted
k - number of return parameters: returns depend on the operation performed

Python Example using PyMongo

PyMongo is a MongoDB library for Python.

"""
mongodb_using_pymongo.py
performs a database update
"""

from pymongo import MongoClient

# Define parameters
host = "localhost"
port = 27017
name = "ml_database"
collection = "customers"
query = {"customer": "1"}
data = {"first_name": "john", "last_name": "jones"}

# Open the database.
connection = MongoClient(host=host, post=port)
database = connection(name)
database_collection = database(collection)

# Update a document.
result = database_collection.update_one(query, data)

# Close the database.
connection.close

Commercial Products

Commercially available products include: