Knowledge Analytics Automation Scripts with SQL Saved Procedures

Picture by Editor

# Introduction

 
Knowledge has grow to be a better commodity to retailer within the present digital period. With the benefit of getting ample knowledge for enterprise, analyzing knowledge to assist firms acquire perception has grow to be extra crucial than ever.

In most companies, knowledge is saved inside a structured database, and SQL is used to amass it. With SQL, we are able to question knowledge within the kind we would like, so long as the script is legitimate.

The issue is that, generally, the question to amass the information we would like is complicated and never dynamic. On this case, we are able to use SQL saved procedures to streamline tedious scripts into easy callables.

This text discusses creating knowledge analytics automation scripts with SQL saved procedures.

Curious? Right here’s how.

# SQL Saved Procedures

 
SQL saved procedures are a set of SQL queries saved straight throughout the database. In case you are adept in Python, you’ll be able to consider them as features: they encapsulate a sequence of operations right into a single executable unit that we are able to name anytime. It’s helpful as a result of we are able to make it dynamic.

That’s why it’s useful to grasp SQL saved procedures, which allow us to simplify code and automate repetitive duties.

Let’s attempt it out with an instance. On this tutorial, I’ll use MySQL for the database and stock data from Kaggle for the desk instance. Arrange MySQL Workbench in your native machine and create a schema the place we are able to retailer the desk. In my instance, I created a database known as finance_db with a desk known as stock_data.

We are able to question the information utilizing one thing like the next.

USE finance_db;

SELECT * FROM stock_data;

Normally, a saved process has the next construction.

DELIMITER $$
CREATE PROCEDURE procedure_name(param_1, param_2, . . ., param_n)
BEGIN
    instruct_1;
    instruct_2;
    . . .
    instruct_n;
END $$
DELIMITER ;

As you’ll be able to see, the saved process can obtain parameters which might be handed into our question.

Let’s look at an precise implementation. For instance, we are able to create a saved process to combination inventory metrics for a particular date vary.

USE finance_db;
DELIMITER $$
CREATE PROCEDURE AggregateStockMetrics(
    IN p_StartDate DATE,
    IN p_EndDate DATE
)
BEGIN
    SELECT
        COUNT(*) AS TradingDays,
        AVG(Shut) AS AvgClose,
        MIN(Low) AS MinLow,
        MAX(Excessive) AS MaxHigh,
        SUM(Quantity) AS TotalVolume
    FROM stock_data
    WHERE 
        (p_StartDate IS NULL OR Date >= p_StartDate)
      AND (p_EndDate IS NULL OR Date <= p_EndDate);
END $$
DELIMITER ;

Within the question above, we created the saved process named AggregateStockMetrics. This process accepts a begin date and finish date as parameters. The parameters are then used as situations to filter the information.

You may name the saved process like this:

CALL AggregateStockMetrics('2015-01-01', '2015-12-31');

The process will execute with the parameters we cross. For the reason that saved process is saved within the database, you should utilize it from any script that connects to the database containing the process.

With saved procedures, we are able to simply reuse logic in different environments. For instance, I’ll name the process from Python utilizing the MySQL connector.

To do this, first set up the library:

pip set up mysql-connector-python

Then, create a perform that connects to the database, calls the saved process, retrieves the outcome, and closes the connection.

import mysql.connector

def call_aggregate_stock_metrics(start_date, end_date):
    cnx = mysql.connector.join(
        consumer="your_username",
        password='your_password',
        host="localhost",
        database="finance_db"
    )
    cursor = cnx.cursor()
    attempt:
        cursor.callproc('AggregateStockMetrics', [start_date, end_date])
        outcomes = []
        for lead to cursor.stored_results():
            outcomes.prolong(outcome.fetchall())
        return outcomes
    lastly:
        cursor.shut()
        cnx.shut()

The outcome can be just like the output under.

[(39, 2058.875660431691, 1993.260009765625, 2104.27001953125, 140137260000.0)]

That’s all you must find out about SQL saved procedures. You may prolong this additional for automation utilizing a scheduler in your pipeline.

# Wrapping Up

 
SQL saved procedures present a technique to encapsulate complicated queries into dynamic, single-unit features that may be reused for repetitive knowledge analytics duties. The procedures are saved throughout the database and are simple to make use of from totally different scripts or purposes resembling Python.

I hope this has helped!
 
 

Cornellius Yudha Wijaya is a knowledge science assistant supervisor and knowledge author. Whereas working full-time at Allianz Indonesia, he likes to share Python and knowledge ideas by way of social media and writing media. Cornellius writes on quite a lot of AI and machine studying subjects.