Question

I have some web application for placement on external sites. This application is a widget for comments like disqus (disqus.com).

In application each widget it’s a entity object with a lot of many attributes.
For instance:

@Entity
class Widget {

    private User admin;        //same as owner, administrator, creator

    private String site;       //example.com

    private String providers;  //google_plus,facebook,twitter,lastfm and others

    //... and a lot of many other attributes
}

At the now I develop the list of the bad words. This list will be managed by an administrator widget (owner).

For this task I have two ways for implements:

First way:

@Entity
class Widget {

    ...
    @ElementCollection(fetch = FetchType.LAZY)
    @Column(name = "bad_words", length = 20)
    private Set<String> badWords= new HashSet<String>();
    ...
}

that is keep as relation in second table

Second way:

@Entity
class Widget {

    ...
    @Column(name = "bad_words", length = 2000)
    private String badWords;
    ...
}

that is keep as string with words splited by some separator, for instance: “fukc,azz,shitt,…”

So the question is what’s the best way to keep bad words in the each widgets entity? Case 1, 2 or some thing else?

Answer 1

Definitely Option 1. Option 2 violates First Normal Form.

Think about having all the words concatenated as discuss in the Second Way. A lot of difficulties in querying the table will arise. Also, a lot of awful code for updating/deleting/adding words will appear. You don’t want any of this.

Having all the words in a collection (First Way) will permit querying them easily and will make all the CRUD operations simpler.