SSMS SQL Row_Number incorrect sort - sql-server

Question

I'm trying to sort by USID and then Date and restart the row number on USID, Utilized. But for some reason the Date is in the incorrect order, so the row numbering isn't doing what I need.
Currently this is what my code is getting me:
USID | Date | Utilized | RowNumber
123 | 1/1/19 | T | 1
123 | 1/2/19 | T | 2
123 | 1/3/19 | T | 3
123 | 1/19/19| T | 4
123 | 1/20/19| T | 5
123 | 1/21/19| T | 6
123 | 1/4/19 | F | 1
123 | 1/5/19 | F | 2
123 | 1/6/19 | F | 3
123 | 1/7/19 | F | 4
123 | 1/8/19 | F | 5
123 | 1/9/19 | F | 6
125 | 1/1/19 | T | 1
125 | 1/2/19 | T | 2
125 | 1/3/19 | T | 3
125 | 1/10/19| T | 4
125 | 1/11/19| T | 5
125 | 1/12/19| T | 6
125 | 1/4/19 | F | 1
125 | 1/5/19 | F | 2
125 | 1/6/19 | F | 3
125 | 1/7/19 | F | 4
125 | 1/8/19 | F | 5
125 | 1/9/19 | F | 6
However I would want it to look like:
USID | Date | Utilized | RowNumber
123 | 1/1/19 | T | 1
123 | 1/2/19 | T | 2
123 | 1/3/19 | T | 3
123 | 1/4/19 | F | 1
123 | 1/5/19 | F | 2
123 | 1/6/19 | F | 3
123 | 1/7/19 | F | 4
123 | 1/8/19 | F | 5
123 | 1/9/19 | F | 6
123 | 1/19/19| T | 1
123 | 1/20/19| T | 2
123 | 1/21/19| T | 3
125 | 1/1/19 | T | 1
125 | 1/2/19 | T | 2
125 | 1/3/19 | T | 3
125 | 1/4/19 | F | 1
125 | 1/5/19 | F | 2
125 | 1/6/19 | F | 3
125 | 1/7/19 | F | 4
125 | 1/8/19 | F | 5
125 | 1/9/19 | F | 6
125 | 1/10/19| T | 1
125 | 1/11/19| T | 2
125 | 1/12/19| T | 3
Here is what I have so far:
select
USID, cast(Date as datetime) as Date, Utilized,
ROW_NUMBER() OVER(PARTITION BY USID, Utilized
ORDER BY USID, cast(Date as datetime) ASC, Utilized)
from vps_time_FullyUtilized
Where am I going wrong?

Answer 1

Here is one gaps and islands solution.
SQL Fiddle
MS SQL Server 2017 Schema Setup:
CREATE TABLE T(USID INT, Date DATETIME, Utilized NVARCHAR(10))
INSERT INTO T VALUES
(123, '1/1/19', 'T'),
(123, '1/2/19', 'T'),
(123, '1/3/19', 'T'),
(123, '1/4/19', 'F'),
(123, '1/5/19', 'F'),
(123, '1/6/19', 'F'),
(123, '1/7/19', 'F'),
(123, '1/8/19', 'F'),
(123, '1/9/19', 'F'),
(123, '1/19/19', 'T'),
(123, '1/20/19', 'T'),
(123, '1/21/19', 'T'),
(223, '1/1/19', 'T'),
(223, '1/2/19', 'T'),
(223, '1/3/19', 'T'),
(223, '1/4/19', 'F'),
(223, '1/5/19', 'F'),
(223, '1/6/19', 'F'),
(223, '1/7/19', 'F'),
(223, '1/8/19', 'F'),
(223, '1/9/19', 'F'),
(223, '1/19/19', 'T'),
(223, '1/20/19', 'T'),
(223, '1/21/19', 'T')
Query 1:
WITH DataWithEndBoundries AS
(
SELECT
USID,Date,Utilized,
--Mark your groups here
IsNewGroup = CASE WHEN ISNULL(LAG(Utilized) OVER (ORDER BY USID,Date),Utilized)<>Utilized THEN 1 ELSE 0 END
FROM
T
)
,VirtualGroup AS
(
SELECT
*,
--This serialzes the marked groups into clusters
VirtualGroupID = SUM(IsNewGroup) OVER (ORDER BY USID, Date ROWS UNBOUNDED PRECEDING)
FROM
DataWithEndBoundries
)
SELECT
USID, Date, Utilized,
--Now you can row number agianst the Virtualized value
RowNumber = ROW_NUMBER() OVER (PARTITION BY USID, VirtualGroupID ORDER BY Date)
FROM
VirtualGroup
ORDER BY
USID,Date
Results:
| USID | Date | Utilized | RowNumber |
|------|----------------------|----------|-----------|
| 123 | 2019-01-01T00:00:00Z | T | 1 |
| 123 | 2019-01-02T00:00:00Z | T | 2 |
| 123 | 2019-01-03T00:00:00Z | T | 3 |
| 123 | 2019-01-04T00:00:00Z | F | 1 |
| 123 | 2019-01-05T00:00:00Z | F | 2 |
| 123 | 2019-01-06T00:00:00Z | F | 3 |
| 123 | 2019-01-07T00:00:00Z | F | 4 |
| 123 | 2019-01-08T00:00:00Z | F | 5 |
| 123 | 2019-01-09T00:00:00Z | F | 6 |
| 123 | 2019-01-19T00:00:00Z | T | 1 |
| 123 | 2019-01-20T00:00:00Z | T | 2 |
| 123 | 2019-01-21T00:00:00Z | T | 3 |
| 223 | 2019-01-01T00:00:00Z | T | 1 |
| 223 | 2019-01-02T00:00:00Z | T | 2 |
| 223 | 2019-01-03T00:00:00Z | T | 3 |
| 223 | 2019-01-04T00:00:00Z | F | 1 |
| 223 | 2019-01-05T00:00:00Z | F | 2 |
| 223 | 2019-01-06T00:00:00Z | F | 3 |
| 223 | 2019-01-07T00:00:00Z | F | 4 |
| 223 | 2019-01-08T00:00:00Z | F | 5 |
| 223 | 2019-01-09T00:00:00Z | F | 6 |
| 223 | 2019-01-19T00:00:00Z | T | 1 |
| 223 | 2019-01-20T00:00:00Z | T | 2 |
| 223 | 2019-01-21T00:00:00Z | T | 3 |