Architecture

Architecture of PQL is separated in following segments:

• Query
• Driver Query Builder
• Driver
• Result Builder
• Query Runner

Query

This is a general abstraction of the query that the user will be using. It supports all common keywords and elements that a specific database type supports.

SQL Queries

For queries specific to SQL databases, they will be in Sql/Query namespace.

User creates this query by using various helper functions like select(), insert(), update(), delete() or their class counter parts Select, Insert, Update, Delete, each of these methods or classes support method chaining for adding conditions, joins, limits, offsets, etc.

All of these inherit from the same StructuredQuery interface and can be combined between in various ways opening up many different approaches for any kind of complex query.

Driver Query Builder

This query builder takes a Query as an input and returns a Raw Query as an output.

Builder creates a driver specific representation of the query which can be a string in case of a SQL builder or different results like arrays depending on what a specific driver needs.

Driver

This is the main executor of the driver specific query. It accepts a Raw query and performs all needed logic for instantiation, like connecting to the database, providing credentials, setting up configurations etc.

This driver will pass the raw query into the database as a request and return the resulting data from the database

Result Builder

Results from the driver are passed into result builder. Result builder allows for a final preprocessing step before the data is actually returned. User can use a result builder to set a pipeline for processing specific results and create things like ORM.

Query Runner

Query runner combines a Driver, Driver Query Builder and a Result Builder into a simple-to-use package so that the user can instantiate everything once, pass it into Query Runner and then just process queries through Query Runner. This abstraction provides different ways of retrieving data from taking one record, multiple, or to process it through a result builder.