Metadata/Forms of Metadata
Descriptive metadata
[edit | edit source]Structural metadata
[edit | edit source]Certain communities, such as database administrators, are more concerned with another type of metadata: structural metadata. This type of metadata documents how data is structured and stored.
For example, a store might keep a database that tracks how much money customers spend on groceries. In this case, we would know that the customer's name would always be a string of characters, and that the amount of money they spend would always be a positive number. Knowing this, we would be able to set up a database that stores the data more efficiently. While creating the database, they might issue a command like:
CREATE TABLE purchases(
purchase_id UNSIGNED INTEGER,
customer_name VARCHAR(50),
amount_spent UNSIGNED INTEGER,
purchase_date TIMESTAMP,
PRIMARY KEY (customer_id)
);
This establishes a schema -- a set of constraints -- for our database table. Here are some of the constraints:
- The
purchase_date
must be in SQL'stimestamp
. The database will throw an error if somebody tries to enterLast Tuesday
for a transaction'spurchase_date
, but will accept2020-02-14 09:22:15
. - Customers cannot spend a negative amount of money during a purchase, because the
amount_spent
column only allows values with anUNSIGNED INTEGER
type, which must be positive. - Customers with very long names may find their names shortened in the database, since the
VARCHAR(50)
type only allows 50 characters of data in thecustomer_name
column.
After the database has seen some use, we might have the following in our table:
purchase_id (UNSIGNED INTEGER) | customer_name (VARCHAR(50)) | amount_spent (UNSIGNED INTEGER) | purchase_date (TIMESTAMP) |
---|---|---|---|
1 | Ashok Kumar | 2784 | 2020-02-14 09:22:15 |
2 | Hong Gil dong | 3516 | 2020-02-14 09:30:02 |
3 | Joe Borg | 499 | 2020-02-14 09:31:43 |
4 | Israel Israeli | 2495 | 2020-02-15 09:00:04 |
By themselves, the values 499
, Hong Gil dong
, 4
, and 2020-02-14 09:22:15
are meaningless. However, the structural metadata we entered allows us to define, interpret, maintain, and efficiently store the data in the table.
Confusingly, structural metadata has a second definition. In the digital library community, the term structural metadata is frequently used to refer to a specific type of descriptive metadata, namely metadata that describes the physical and/or logical structure of a particular communication. For example, if a digital library project had created digital images of all the pages in a book, structural metadata would document the order of the pages, and perhaps which page each chapter or section began with.