snap_events(user_id, event_type, event_at) and users(user_id, age_bucket)

For September 2026, return for each age bucket what percentage of its events were sends and what percentage were opens. Other event types are ignored entirely, including in the denominator.

Return age_bucket, send_pct and open_pct rounded to two decimals, ordered by age_bucket.

The two percentages add to 100 within a bucket, because the denominator is sends plus opens rather than all events.

Tables

snap_events

user_id  event_type  event_at
-------  ----------  -------------------
1        send        2026-09-01 10:00:00
1        send        2026-09-02 10:00:00
1        open        2026-09-03 10:00:00
2        open        2026-09-04 10:00:00
2        send        2026-09-04 11:00:00
1        screenshot  2026-09-05 10:00:00
3        send        2026-09-06 10:00:00
4        open        2026-09-07 10:00:00
... 1 more row(s)

users

user_id  age_bucket
-------  ----------
1        18-24
2        18-24
3        25-34
4        25-34

Expected result

age_bucket  send_pct  open_pct
----------  --------  --------
18-24       60.00     40.00
25-34       50.00     50.00