Implement grouping options for date & time types

[kgodey]

Problem

Users may want to group their records with date and time fields by common intervals such as hours, months, quarters, years, etc. The appropriate grouping for a field will vary depend on the kind of data stored in that field. For example, if there are years ranging from 1500-2000 in the data, the user might want to group dates by century, but if the dates only range from 2008-2010, the user may want to group by month or quarter instead and centuries are useless.

We need to provide both useful grouping intervals for a given column and the ability to group by them.

Solution

• We should support grouping by all valid date parts supported by Postgres' EXTRACT function. Documentation.
• We should provide useful grouping intervals for a given column in the Column API.
• We should extend our record grouping mechanism to accept date parts for columns of Date & Time types. We should only accept valid date parts based on the type of column (e.g. we should not accept century for TIME columns).

This involves:

• Implementing the grouping in the backend
• Updating the /api/v0/databases/<id>/types/ endpoint to store available grouping on this type

Additional context

• Please see these PRs for code related to grouping:
  • Record grouping #315
  • Fix grouping counts #353
• Complete implementation of this issue is blocked by implementation of Date & Time types.
  • Handle TIMESTAMP data type in the backend #424
  • Handle DATE data type in the backend #425
  • Handle TIME data type in the backend #426
• Update filtering issues to cover storing filter/group information in the backend. #557 provides some context on storing groups.