hyp2ue_t2_SQL

This repository provides Exercises for web development environment designed for the use in web development classes at the Upper Austria University of Applied Sciences (FH Oberösterreich), Hagenberg Campus.

Installation

See INSTALL.md

To edit SQL-Statements you can use any text editor, but I would recommend using PHPStorm with its DataGrip plugin. The database is installed in a Virtual Box and maintained with Vagrant or in a Docker Image .

Install or Restore onlineshop database with Docker

docker exec -it mariadb /bin/bash -c "mariadb -uonlineshop -pgeheim </src/onlineshop.sql"

Exercises

Goal of the three lessons is to learn SQL. We start with DDL und focus on DML in the second part. Part three deals with normal forms and ER models.

Besides DataGrip you can use PHPMyAdmin, MySQLWorkbench and many other Tools for database administration to run SQL-Scripts. But PHPStorm provides a full integration of all technologies we need to use SQL in a web application.

Open a PowerShell or a Terminal on MAC and go to the directory, where your Vagrantfile ist stored: cd <path-to>/fhooe-webdev and start the Virtual Box Image with vagrant up.

Open the Projekt hypue2-t2-SQL in PHPStorm. Press the right mouse button and choose run onlineshop.sql. Choose the preconfigured MariaDB-Connection. If that works you can start. Otherwise check your Configuration of the database connection.

Alternatively you can import the script onlineshop.sql with PHPMyAdmin. Open https://192.168.7.7/phpmyadmin in a browser. Login with onlineshop/geheim and go to the tab import.

A third option is to do this in a commandline. With the provided Vagrant image, the steps below are necessary.

vagrant ssh
cd /var/www/html/code/hyp2ue-t2-SQL
sudo bash
mariadb

MariaDB [(none)]> \. onlineshop.sql

Each step of the lessons should be marked with a comment like this:

-- 1

Use multiline comments to answer questions within the SQL file:

/*
 You need the blank '-- ' to make it an SQL comment
 not writing the blank --causes a syntax error
*/

part 1 DDL

Use sql/login.sql as a starting point.

Create a table with an appropriate name

it should contain the columns email, password and active.
Add an appropriate primary key (BIGINT, UNSIGNED and AUTOINCREMENT)
All columns are required fields.
active will contain a MD5 hash.
Use appropriate data types and length
What could be the point of saving this table in an additional database outside onlineshop

Create a table with 6 columns

A primary key with BIGINT UNSIGNED AUTOINCREMENT, a TEXT column, a VARCHAR, a DATE, a DECIMAL and a column, that will become a foreign key.
Choose column names that make sense together with the data type.
All columns are required fields.
Consider adding an index (KEY) to the column that will become foreign key.
Add a foreign key to the chosen column in an additional statement and connect the current table with the one from step 1.

Add 5 rows to each of the tables above.

Use all 3 types of insert to do this.
Use appropriate, fictional values for each column.
Use MD5(RAND(1)) for the column active to hash the value of a randomized seed.
Look into MariaDB documentation to find possible encryption functions for the password. Do you think MariaDB provides functions, that are secure enough? Which functions does PHP provide to encrypt and verify passwords?

Delete the content of the tables in the right order

PK, FK order?
Use two different kind of statements to do this.

Delete the tables itself in the right order
Do some research on "dynamic columns"

Consider which tables of the onlineshop schema are candidates to use dynamic columns.
Create an example table, that makes use of dynamic columns. Not the one from the MariaDB documentation! That will give 0 points for this step.
JSON is already used here. Read the data in JSON format. Where does JSON come from?
Delete the table created for the dynamic column example.

The Script created for this exercise should run without errors. For example:

user@ubuntu1/mycode$ sudo mysql –uonlineshop –pgeheim <login.sql

item_name COLUMN_JSON(dynamic_cols)
MariaDB T-shirt {"size":"XL","color":"blue"}
Thinkpad Laptop {"color":"black","price":500}

But you can use DataGrip/PHPStorm or PHPMyAdmin as well.

part 2 DML

Use sql/ue2.sql as a starting point. See if you can run the script with clicking the right mouse button and choose run ue2.sql. Otherwise check your Configuration of the database connection.

If you need content for inserting, updating or comparing values look into the tables already provided in the onlineshop schema. For example: insert into user set email = 'shopuser1@onlineshop.at'. Although this is commonly used you do not have to capitalize reserved words. Syntax highlighting in modern editors has the same effect.

Write an SQL statement to insert a data row in onlineshop.user

Make use of the AUTOINCREMENT on the column iduser.
Use now() for the date value.
For the column active use MD5(RAND(<number> | <string>)). This hash is used for two-phase authentification. Therefore an email is sent to the user with a link containing the hash value stored in active: www.maindomain.org?active=hash. If the user clicks this link a program on the server compares the sent hash with the database and sets active to null. Only then the authentification is valid and the registration finished.
The column role has a default value. If you need other values like 'admin' you have to add this column to the insert statement.
Forpassword you can use a plain string. Later we will use a hash value generated by the PHP function password_hash().

Write an SQL statement that checks if a combination of a password and an email is already stored in the database, and if the account is already activated.

In the select clause use iduser, first_name, last_name and nick_name.
Compare onlineshop.user.password with a plain string. Embedded in PHP we will use the function password_verify() to do this. Then password has to be part of the select clause.
active has to be compared as follows: active is null
These are the three constraints for a valid login to OnlineShop.

Set an appropriate value for onlineshop.product.product_category_name for all products already stored.

Which kind of statement can do the job?
Have a look at onlineshop.product_category.product_category_name to find appropriate values.
Write 5 or 2 Statements to change the values.
You need not bring it down to one statement using subselects and wildcards or case.

Write a statement that adds a new data row to onlineshop.product.

product_category_name should be 'Wohnung'.
Make use of the AUTOINCREMENT on idproduct.
Use now() for the date.
Set active to 1. That means the product should be listed on the website.
For all other columns use values that make sense.

Write an SQL statement that copies data from onlineshop.cart to onlineshop.orders and onlineshop.order_item. This two statements build the step final checkout or order with costs.

See SQL_exercise_part2.pdf for an ER-Modell.
Use two types of subselects for this.
Which insert has to be the first one, considering thatorder_item.orders_idorders is a foreign key pointing at orders.idorders?
For one order always group data rows together that have the same session_id.
Instead of a value for total_sum use a subselect on the right position within the statement. insert into orders ( ..., total_sum, ...) values ( ..., select that returns only one value, ...)
For order_item use a subselect to transfer all entries from cart. For each column in the table description of order_item set an appropriate column in the select clause in the right order.
cart.orders_idorders can be filled with last_insert_id().
Consider, that you will need on statement for each session_id
cart is an inMemory table. Therefore it is emptied at every reboot. To get test data, see onlineshop.sql

Write a statement that lists all turnovers grouped by product category.

Use a join between order_item and product to handle this.
This is a step towards data warehouse. Data warehouses are database instances that are optimized for read performance and data aggregation jobs.
Add a having clause to limit the result set to data rows with a turnover less than 100000 €.

To test your script run onlineshop.sql before ue2.sql. Both scripts must not return errors. You can use PHPStorm/DataGrip and run for this. You can use PHPMyAdmin/Import or the commandline as well.

cd <path-to-onlineshop.sql>
sudo bash
mariadb
[MariaDB (none)] \. onlineshop.sql
[MariaDB (none)] \. ue2.sql

part 3 normal forms and ER models

Create a new file with an arbitrary name <mysolutionfilename>.sql in the sql folder. The extension has to be .sql. This file should transfer the original data model created by onlineshop.sql into the data model given by the steps of this exercise.

To generate a visualization of the data model goto PHPStorms DataGrip Widget after running onlineshop.sql. PHPStorm ER Diagram Visualization

Click Database in the right upper corner
Right click on onlineshop
Click on Diagrams at the end of the list
Click Show Visualization

There are two ways, how you can handle this.

Copy onlineshop.sql and alter the original statements in a way that the new data model is created.
Run onlineshop.sql and afterwards run the new script containing alter table statements that change the physical model.

The former script is used to build a new installation script for a new digital product or a new version of it. The later script is used to change a data model already used in production to add new features or delete unnecessary features.

All tables are in 1NF unless otherwise stated.

In this exercise we will focus on the second scenario with alter table statements.

Have a look at the data model and see which tables already meet 2NF or not.

Which tables meet 2NF at the first glance? Give a rationale for this.
Which tables need a closer look to decide on this?
Give a rationale for each of these tables, why they meet 2NF or not.

Have a look at the tables address, city and country. Explain, why the CHASM trap exists in this case or why it does not. A relevant question is a query for all address fields.
Have a look at the tables order, user and visit. Explain, why the FAN trap exists in this case or why it does not. A relevant question is a query that asks at which visit the order was placed.
Look at the table address. Does it meet 1NF, if it is necessary to ask for street, house number, door number and floor directly? Change the table address in a way that it meets this new requirements.
address meets 3NF in both cases, because the columns of city and country are in separate tables. Change address in a way that querying the whole address becomes faster. city and country should remain as lookup tables. Give reasons for using lookup tables and how they are used in web development.
Change the tables product and product_category in a way that they meet 3NF.
Look at the table payment. Three columns do not have speaking names. Add key columns with alter table to provide a column to store speaking names for the values. Additionally think of a solution to store as many attributes and values for different payment variants as necessary. Implement the variant discussed in UE1 as payment_v2. The payment_v2 table should be linked to the table user in a proper way.
Rework the table order in a way that the columns for payment match one of the 3 variants of 7.
Create a new table user_visit based on the columns of user and visit (CREATE TABLE .. AS SELECT + JOIN). Which problems will you face, if the tables user and visit would no longer exist? What happens, when user login. How do people change their password, emails, try to login or register? How is this stored in this case? What information can be queried from the new table (append only instead of updates), that has not been there, when the information was stored in user and updated later.

Digital-Media / hyp2ue_t2_SQL