Getting total using Union All - mysql

Question

So here is my situation. I have 1 query that gets results grouped by date then I union all a second query that gets the totals (no grouping by date). My issue is I am calculating the average of fields and when I want to total up the average my numbers don't add up.
Here is my SQLFiddle
Here is my query:
SELECT
t.end,
SUM(CASE WHEN (t.start != t.end) THEN TIMESTAMPDIFF(DAY, t.start, t.end) ELSE 1 END) / COUNT(t.id) as averageTime
FROM store t
GROUP BY t.end
UNION ALL
SELECT
'Total',
SUM(CASE WHEN (t.start != t.end) THEN TIMESTAMPDIFF(DAY, t.start, t.end) ELSE 1 END) as averageTime
FROM store t
Right now the second query just gives the total not the total of the average. Any help is appreciated, thank you.
To clarify since there is some confusion...
I need to get the average per grouped date by how many items were in that group
timeDiff / count(t.id)
Since I am not grouping in the union all query it is doing it as a whole then dividing. I hope that makes more sense.
The first query is correct the data output is as follows:
1, 1.6667, 3 (Those are the averageTime values from the first query)
5.6667 (Should be the total row) Right now I have it out putting 10 that is the total before the first rows were averaged out.

Answer 1

THe first part of the query can probably be written as:
SELECT t.end,
AVG(CASE WHEN t.start != t.end THEN TIMESTAMPDIFF(DAY, t.start, t.end) ELSE 1
END) as averageTime
FROM store t
GROUP BY t.end;
Presumably, you want the averages of the averages -- rather than the overall average. I assume this because your query is calculating the overall average.
One way is the brute force way:
SELECT t.end,
AVG(CASE WHEN t.start != t.end THEN TIMESTAMPDIFF(DAY, t.start, t.end) ELSE 1
END) as averageTime
FROM store t
GROUP BY t.end
UNION ALL
SELECT 'Total', avg(AverageTime)
FROM (SELECT t.end,
SUM(CASE WHEN t.start != t.end THEN TIMESTAMPDIFF(DAY, t.start, t.end) ELSE 1
END) as averageTime
FROM store t
GROUP BY t.end
) t;

Answer 2

Your question is not clear. But, I suppose you need Average. As #samD mentioned,
the second part of you query can be written as
SELECT
'Total',
AVG(CASE WHEN (t.start != t.end) THEN TIMESTAMPDIFF(DAY, t.start, t.end) ELSE 1 END) as averageTime
FROM store t

Answer 3

You're mistakenly assuming that the COUNT(t.id) in the first query represents the total number of rows, whilst it is actually also affected by the GROUP BY clause. If you multiply the separate rows by their number of days the totals also add up. Remove the division by COUNT(t.id) and the totals will add up, and you will correctly have the grouped SUMs as expected.
I'm just not sure how that solves your problem since you're not really clear on that: you simply cannot add all averages together and expect them to add up to a global average, that's not how math works: the sum of averages of arbitrary sized subgroups is rarely equal to the straight unweighted average of all groups together.