CODEWITHSUNDEEP
database-designblob
jl6
jl6

Reputation: 6394

How do I design a database for information that does not YET have a schema?

Say I want to build a database of emails. As I do not understand the structure of an email, I might store emails in a simple table with only basic attributes:

Emails
   email_id   integer (PK)
   from       varchar(200)
   sent       timestamp
   email      blob

But then later I decide that I understand the structure of an email after all and want to extend my database to represent the whole email as structured data instead of just a blob. I would probably create an EmailAddress entity, a MIME entity, a Header entity, with appropriate foreign keys and associative tables.

But what can I do in my database now in anticipation of one day unpacking a blob into a subschema?

Upvotes: 1

Views: 92

Answers (3)

Damir Sudarevic
Damir Sudarevic

Reputation: 22187

  • Mask future schema changes by implementing an access layer. Use views for reading, stored procedures (or triggers) for writing to underlying tables -- avoid direct access from app layer to tables.

  • Consider anchor modelling approach, which uses 6NF and is quite suitable for constant changes -- schema evolution. The anchor modelling toll automatically creates views and triggers as an access layer.

Upvotes: 0

Phil Sandler
Phil Sandler

Reputation: 28016

Consider storing it in a structured format (e.g. XML) so that you can effectively convert the data to a proper format later.

Upvotes: 1

Oded
Oded

Reputation: 499002

Don't use a relational database for this.

Use a document database, or some other schemaless data store.

Upvotes: 1

Related Questions