Do you prefer verbose naming when it comes to database columns? [closed]

Answer 1

ID is terribly confusing when you have lots of joins. I prefer to have explict names of ids that match the name of the id in the foregin key. That way you always know that customerid will join with customerid in any other table that has the column.

Never use name as a field name. It is a reserved word and thus should be avoided. Reserved keyworsds are used by the database in some fashion, using them also as filed names is just poor proactice and can create more errors than using the correct descriptive name form the start.

Answer 2

I'm a fan of option 3:

CREATE TABLE Products
(
    ProductId int NOT NULL IDENTITY(1,1) PRIMARY KEY,
    CategoryId int NOT NULL FOREIGN KEY REFERENCES Categories(CategoryId),
    Name varchar(200) NOT NULL
)

So the primary key is the only column to gain the table's name as a prefix -- IMHO it makes it easier to see when a join has gone wrong. Then again, I also like using GUIDs for primary keys if there is any possibility of having to cope with a merge replication situation at any point in the future...

Answer 3

Check also related question: Is prefixing each field name in a table with abbreviated table name a good practice?

As I mentioned in my answer; the concept of prefixing field names with the table name comes from the old time of legacy systems when each field across the whole database needed to be unique. That is not required any more by the modern systems so it is just a convention that is no longer necessary. As mentioned by Think Before Coding "The table name already gives the context. No need to prefix columns name with it"

Answer 4

I prefer uid, gid, pid, wid, lid, etc..

Answer 5

I'm usually giving each table a unique prefix. So your example would be something like

CREATE TABLE Products
(
    PROD_ID int NOT NULL IDENTITY(1,1) PRIMARY KEY,
    PROD_CAT_ID int NOT NULL FOREIGN KEY REFERENCES Categories(CAT_ID),
    PROD_NAME varchar(200) NOT NULL
)

This makes joining and selecting columns easier, because you have no name conflicts. Unless you reference the same table more than once you will not even need table name aliases (most likely).

However lately I'm starting to thing that (2) might be better, because it's closer to the naming conventions I'm using when writing code (C# in my case).

Answer 6

One school of thought is that a column name should represent a single domain space across the database. For example, a product name is something quite different from a company name. One might be longer than the other, etc. In that respect you should use the first method since it distinguishes between the two.

On the other hand, I agree to some degree with Think Before Coding's logic. I always use the table.column syntax, so the context should be clear.

Using the more verbose column names usually avoids conflicts with reserved words.

Then again, it can be quite redundant if every column in your product table is ProductXxxx.

In other words, I don't have an absolute preference, but I do tend more towards the verbose naming convention.