Ms Access query count first - sql

Question

I have the following table:
tblOrder
----------------------------
OrderID ==>Primary Key
CustomerID ==>Number
OrderDate ==>Date
Order ==>Number
Price ==>Currency
[Order] field contain a number (like 11, 15, 17) which is a code.
I am trying to find a way, and if it is possible of course, to count [Order] per [CustomerID] that:
==> I) Placed only [Order]=15
==> ii) Made first [Order]=15 and then placed another order which IS NOT 15.
Example Table:
OrderID CustomerID OrderDate Order Price
--------------------------------------------
1 3 1/1/2018 15 100
2 2 3/2/2018 15 300
3 2 7/3/2018 11 400
4 3 2/6/2018 15 200
5 1 5/2/2018 17 300
6 1 11/7/2018 15 600
7 2 1/4/2018 11 200
OutPut Query:
CustomerID OnlyOrder=15 FirstOrder=15
---------------------------------------
1 Null Null
2 1 3
3 2 Null
Thank you in advance.

Answer 1

This is much easier using aggregation and having. The following gets the customers:
select customerId
from tblOrders
group by customerId
having min(iif([order] = 15, orderdate, null)) < max(iif([order] <> 15, orderdate, null)) or
(min([order]) = 15 and max(order = 15));
To count them, use a subquery:
select count(*)
from (select customerId
from tblOrders
group by customerId
having min(iif([order] = 15, orderdate, null)) < max(iif([order] <> 15, orderdate, null)) or
(min([order]) = 15 and max([order]) = 15)
) as c;
Edit:
If you want the first order to be 15 --which is not how I interpret the description -- that is also easy:
select customerId
from tblOrders
group by customerId
having min(iif([order] = 15, orderdate, null)) = min(orderdate) or
(min([order]) = 15 and max(order = 15));
You can use this to count the customer ids.
EDIT:
For the revised question, you want something like this:
select customerId,
iif(min([order]) = 15 and max(order = 15), count(*), 0) as all_first_15,
iif(min(iif([order] = 15, orderdate, null)) = min(orderdate) and min([order]) <> max([order]), count(*), 0) as first_order_15
from tblOrders
group by customerId
having min(iif([order] = 15, orderdate, null)) = min(orderdate) or
(min([order]) = 15 and max(order = 15));

Answer 2

Yup, that's quite a set of rules, but certainly possible.
Let's start with a subquery:
Select all customers where the first order (the order with the lowest (minimal) OrderID) is 15:
SELECT t.CustomerID
FROM tblOrder t
WHERE t.Order = 15
AND EXISTS (
SELECT 1
FROM tblOrder i
WHERE t.CustomerID = i.CustomerID
HAVING Min(i.OrderID) = t.OrderID
)
Then, we can use that to count all orders for these users:
SELECT Count(OrderID)
FROM tblOrder c
WHERE EXISTS (
SELECT 1
FROM tblOrder t
WHERE t.Order = 15
AND EXISTS (
SELECT 1
FROM tblOrder i
WHERE t.CustomerID = i.CustomerID
HAVING Min(i.OrderID) = t.OrderID
)
AND t.CustomerID = c.CustomerID
)
GROUP BY c.CustomerID
Or, if you want to split out counts by 15 and not 15:
SELECT DISTINCT o.CustomerID, (SELECT COUNT(s2.OrderID) FROM tblOrder s2 WHERE s2.CustomerID = o.CustomerID AND s2.Order = 15) As Only15Count, (SELECT COUNT(s1.OrderID) FROM tblOrder s1 WHERE s1.CustomerID = o.CustomerID AND EXISTS (SELECT 1 FROM tblOrder s3 WHERE s3.CustomerID = s1.CustomerID AND s3.Order <> 15)) As TotalCount
FROM tblOrders o
WHERE EXISTS (
SELECT 1
FROM tblOrder t
WHERE t.Order = 15
AND EXISTS (
SELECT 1
FROM tblOrder i
WHERE t.CustomerID = i.CustomerID
HAVING Min(i.OrderID) = t.OrderID
)
AND t.CustomerID = o.CustomerID
)
Note that, in contrast to your expected output, nothing will get returned if the first order isn't 15.