SQL to find average of dates - sql-server

Question

I have a table that stores the dates of events, and I would like to find the average number of events by the day of the week.
e.g.,
ID
date
1
2021-09-01
2
2021-09-01
3
2021-09-02
4
2021-09-03
Result:
Wed: 2
Thu: 1
Fri: 1
UPDATE
I am told that I gave a poor example. Here is another attempt:
There was 1000 events in the month of June. Each event is a row with an ID and a date. There can be multiple events in a day (In the table above, imagine that there are 250 events on 9/1 and 150 on 9/2 and 300 on 9/3, etc.). I would like to know the average number of events for each day (Monday, Tuesday, Wednesday, etc.) for the entire month of June
Yes, I can simply divide by four to find the average for June but I am using a single month as an example. In reality, I am having to average out over several months and would like a solution where I am not having to determine the number of weeks between dates manually (e.g., using a date time calculator or counting the weeks with a calendar).
Here is a query I tried
SELECT COUNT(DATEPART(DD, EVENT_DATE) AS 'Event Total', AVG(COUNT(DATEPART(DD, EVENT_DATE)) AS 'Average'
FROM TABLE_1
GROUP BY DATEPART(DD, EVENT_DATE)
I'm having difficulty grouping by the days and then finding the average of the grouping. Is this possible with just t-sql?

Answer 1

Use Day of Week to get your grouping and count(*) to count all occurrences for the grouping
declare #Tmp as table(ID int, date date)
Insert into #Tmp
(ID, date)
values(1,'2021-09-01')
,(2, '2021-09-01')
,(3, '2021-09-02')
,(4, '2021-09-03')
select count(*) freq,DatePart(WEEKDAY,date) DayOfWeek from #Tmp group by DatePart(WEEKDAY,date)

Answer 2

The details of the averaging are still somewhat vague, e.g. should days with no events in June be counted, but this should point you in a suitable direction. dbfiddle.
declare #SampleEvents as Table ( Id Int Identity, EventDate Date );
insert into #SampleEvents ( EventDate ) values
( '20210901' ),
( '20210902' ), ( '20210902' ),
( '20210903' ), ( '20210903' ), ( '20210903' ),
( '20210904' ), ( '20210904' ), ( '20210904' ), ( '20210904' ),
( '20210905' ), ( '20210905' ), ( '20210905' ), ( '20210905' ), ( '20210905' ),
( '20210906' ),
( '20210907' ),
( '20210908' ), ( '20210908' ),
( '20210909' ), ( '20210909' ), ( '20210909' ),
( '20210910' ), ( '20210910' ), ( '20210910' ), ( '20210910' ),
( '20210911' ), ( '20210911' ), ( '20210911' ), ( '20210911' ), ( '20210911' ),
( '20210912' ),
( '20210913' ),
-- ( '20210914' ),
( '20210915' ),
( '20210916' ),
( '20210917' ),
( '20210918' ),
( '20210919' ),
( '20210920' ),
-- ( '20210921' ),
( '20210922' ),
( '20210923' ),
( '20210924' ),
( '20210925' ),
( '20210926' ),
( '20210927' ),
-- ( '20210928' ),
( '20210929' ),
( '20210930' );
select * from #SampleEvents;
with
SampleEventsWithDoW as (
-- Calculate the day-of-week (DoW) for each date.
-- DoW will always return an integer from 1 to 7 with
-- 1 corresponding to Sunday regardless of the setting of
-- DateFirst or Language .)
select Id, EventDate, ( ##DateFirst + DatePart( weekday, EventDate ) - 1 ) % 7 + 1 as DoW
from #SampleEvents ),
SampleEventsSummarizedByDoW as (
-- Summarize the data by day-of-week.
select DoW, Count(*) as EventCount, Count( distinct EventDate ) as NumberOfDays
from SampleEventsWithDoW
group by DoW )
-- The following commented-out select statements can be used to display the
-- intermediate results.
-- select * from SampleEventsWithDoW;
-- select * from SampleEventsSummarizedByDoW;
-- Average the counts.
-- The cast is used to avoid integer division which would result in
-- values that are probably unacceptable to the OP.
select DoW, Cast( EventCount as Numeric(10,2) ) / NumberOfDays as AverageEventCount,
EventCount, NumberOfDays
from SampleEventsSummarizedByDoW;

Answer 3

Use DATENAME to get the name of the week day, then LEFT to get the left-most three characters, group on this:
SELECT
LEFT(DATENAME(dw, [date]), 3),
COUNT([id])
FROM my_table
GROUP BY
LEFT(DATENAME(dw, [date]), 3);