Storing each individual's email text is very space-inefficient that balloons our DBs. For example, as of creating this ticket, email text occupies approximately 90% of Stanford's database and 95% of Yale's. It would be good to refactor our database setup to store the often-repeated text of emails in a more compressed format and link to the individual information that populates the tags within individual emails in a separate column or table.

Suggestion:

• one table for each batch of emails that gets sent; that table contains the template for the email, the recipients, and the other current metadata (timestamps, etc.)
• separate table whose columns are a foreign key to the previous table, each recipient's userid, and tag values for each of the tags available to admins. The fields would be the values of the tags for the recipient at the time the email is sent.

If we want to recreate the textofemail view for admins, we can do so from these tables, or we can simply show the metadata list in a searchable format, which should answer most questions admins would usually have.

Keywords: e-mail e-mails db databases schema