how to create dynamic data fields in sql database - mysql

Question

There is a requirement that arises to handle dyanmic data fields in database level. Say we have a table called Employee and that table has a name, surname, and contact no fields ( 3 basic fields). So as the application progresses, the requirement is that the database and the application should be able to add (handle) dynamic data fields that can be added with type into the database.
Ex: A user will add data of birth, address field dynamically to the Employee table which has basic 3 fields mainly.
The problem is how to cater to this requirement the optimum way?
there is a picture I have designed tables to cater to this, But I am open for industry-standard optimum way of achieving this without having future problems
Please collaborate with this.

Answer 1

You basically have four options for handling such dynamic fields:
Modify the base table structure whenever a new column is added.
Using JSON to store the values.
Using a EAV model (entity-attribute-model).
Basically (1) but storing the additional values in a separate table or separate table per user.
You have not provided enough information in the question to determine which of these is most appropriate for your data model.
However, here is a quick run-down of strengths and weaknesses:
For modifying the table: On the downside, modifying a table is an expensive operation (especially as the table gets bigger). On the upside, the columns will be visible to all users and have the appropriate type.
For JSON: JSON is quite flexible. However, JSON incurs very large storage overheads because the name of each field is repeated every time it is used. In addition, you don't have a list of all the added fields (unless you maintain that in a separate table).
For EAV: EAV is flexible, but not quite as flexible as JSON. The problem is the value column is a single type (usually a string) or accessing the data gets more complicated. Like JSON, this repeats the "name" of the value every time it is used. However, this is often a key to another table, so the overhead is less.
For a separate table for each user: This primary advantage here is isolating users from each other. If this is a requirement, then this might be the way to go (although adding a userId to the EAV model would also work).
So, the most appropriate method depends on factors, such as:
Will the fields be shared among all users?
Do the additional fields all have the same type?
What are your concerns about performance and data size?
How often will new fields be added?

Answer 2

To have dynamic fields you can use another table where you can set properties of the user
user table has columns
userid, name, surname, contact
user_props table has columns
propertyid, userid, property, value
in user_props you can insert user properties like
INSERT INTO user_props (userid, property, value)
VALUES (1, "date_of_birth", "2010-01-10"),(1, "hobby", "Stackoverflow")
Like this you can dynamically set any number properties to user.

Answer 3

You might be better off using MongoDB or some other NoSQL/Schemaless database which stores your data in key => value pairs. For the fields you are certain about in advance, you can set a type (in MongoDB) so those columns will have a schema. For dynamic fields, the fields would be stored as strings and you would have to figure out the types in your code somehow.
If you need to use MySQL, in your Employee table, you could have a fourth column for custom fields - database type json. Then whenever you add a new custom field, you add the field_name, field_value and field_type. You schema could look like:
//Schema for Employee table in mysql
id: int
name: varchar
surname: varchar
custom_fields: json //eg { {field_name: DOB, field_value: 06/09/2020, field_type: date},... }
//Schema for contacts table
id: int
employee_id: int
contact: varchar
In MySQL, you could also get rid of the type (if you can do without it) in the custom_fields and structure the json to be simple key => value pairs so it looks like
{
{"key":"Age","value":"10"},
{"key":"salary","value":"40,000"},
{"key":"DOB","value":"06/09/2020"},
}
What you seem to be designing here is a variation of the Entity-Attribute-Value model. It works but it would be very cumbersome to query against a schema like that. Using a json column is a lot neater and a lot faster. Best is to use MongoDB and figure out the types in your code.

Answer 4

You may deal with this condition using a stored procedure including an ALTER DATABASE sentence. Something like:
DROP PROCEDURE IF EXISTS set_dynamic_table;
DELIMITER //
CREATE PROCEDURE set_dynamic_table (IN _field_name VARCHAR(50),
IN _field_type VARCHAR(20),
IN _last_field_name varchar(50))
BEGIN
DECLARE _stmt VARCHAR(1024);
SET #SQL := CONCAT('ALTER TABLE dynamic_table',
'ADD COLUMN ' _field_name, field_type, NULL AFTER ,last_field_name);
PREPARE _stmt FROM #SQL;
EXECUTE _stmt;
DEALLOCATE PREPARE _stmt;
END//
DELIMITER;
and you will e