Question

I have two tables

raw_commits(sha,date,author)

and

milestones(name,start,end)

where the milestones table contains the timestamp range for all the project’s milestones.

I am unsure on how to go about the JOIN to achieve the following:

For each raw_commits, show the raw_commits.sha, raw_commits.date, raw_commits.author, milestones.name if it has one milestone, show NULL if it doesn’t “belong” to a milestone, and show duplicate rows for each milestone, if the commit “belongs” to multiple milestones.

The categorization into milestones is obviously done by using the raw_commits.date as a reference, i.e. raw_commits.date >= milestones.start AND raw_commits.date <= milestones.end.

How to achieve the described behavior?

Answer 1

SELECT
  raw_commits.sha,
  raw_commits.author
  raw_commits.date,
  milestones.name
FROM
  raw_commits
LEFT OUTER JOIN
  milestones
ON
  raw_commits.date BETWEEN milestones.start AND milestones.end