Sql Select query - how to order based on an ID - sql-server-2008

Question

I m trying to select records and ordering it using a specific column's value.
Users are assigned to a specific district or area (are associated with districtid and areaid). A user with AreaID of 5 and DistirctID of 0 has edit access to all data within his/her Area which spans multiple districts. A user with AreaID of 5 and DistirctID of 8 has edit access to all data within this district only.
To prevent users from paging through data to find their data, I want to show theirs on top of the table. Don't know how to set the "order by" (order by districtid, but want districtid 8 to be on top). Is it possible?
This is used within an ASP.NET (C#) application.
Assuming I am trying to display data for a user assigned to AreaID 5 and DistrictID 8:
select col1, col2, ..., colN from SomeTable order by DistrictID ... (?)
Database is SQL Server 2016.
Update
this is what i tried:
select distinct ml.MPOOID, ml.MPOOName, ml.AreaID, ml.DistrictID,
zc.AreaName, zc.DistrictName
from MPOO_List ml
left join ZIPCodes zc ON ml.AreaID = zc.AreaID AND ml.DistrictID = zc.DistrictID
where ml.DistrictID=34
union
select distinct ml.MPOOID, ml.MPOOName, ml.AreaID, ml.DistrictID, zc.AreaName, zc.DistrictName
from MPOO_List ml
left join ZIPCodes zc ON ml.AreaID = zc.AreaID AND ml.DistrictID = zc.DistrictID
where ml.DistrictID<>34
if I leave out the left joins, it works.
Would I be able to also order by MPOOName within each group?
Update 2
Using AreaID=5 and DistrictID=34 (for a user who has district access), this works. It shows 516 of 625 records, but the records are ordered as desired: all the ones with AreaID 5 and DistrictID 34 are on top.
To include missing records, I tried changing "AND" to "OR" in second WHERE clause. This gives me the correct 625 records, but no longer in order desired.
Same thing happens if I add additional columns to SELECT. i.e. Changing select to SELECT ml.MPOOID, ml.MPOOName, ml.ManagerFirstName, ml.ManagerLastName, ml.Phone, ml.Email, ml.AreaID, ml.DistrictID, ml.StartDate, zc.AreaName, zc.DistrictName. This returns either 29,196 or 35,191 records, depending on whether I use "AND" or "OR" in second WHERE clause (takes 10 seconds to execute)
SELECT DISTINCT ml.MPOOID, ml.MPOOName, ml.AreaID, ml.DistrictID
FROM MPOO_List ml
LEFT JOIN ZIPCodes zc ON ml.AreaID = zc.AreaID AND ml.DistrictID = zc.DistrictID
WHERE ml.AreaID = 5 AND ml.DistrictID = 34
UNION
SELECT DISTINCT ml.MPOOID, ml.MPOOName, ml.AreaID, ml.DistrictID
FROM MPOO_List ml
LEFT JOIN ZIPCodes zc ON ml.AreaID <> zc.AreaID AND ml.DistrictID <> zc.DistrictID
where ml.AreaID <> 5 AND ml.DistrictID <> 34
actual selected columns needed:
SELECT ml.MPOOID, ml.MPOOName, ml.FirstName, ml.LastName, ml.Phone, ml.Email, ml.AreaID, ml.DistrictID, ml.StartDate, zc.AreaName, zc.DistrictName
Update 3 - Sample DB table and data
CREATE TABLE [MPOO_List](
[MPOOID] [int] NOT NULL,
[MPOOName] [varchar](100) NOT NULL,
[FirstName] [varchar](50) NULL,
[LastName] [varchar](50) NULL,
[Phone] [varchar](14) NULL,
[Email] [varchar](100) NULL,
[AreaID] [int] NULL,
[DistrictID] [int] NULL,
[StartDate] [datetime2](0) NULL,
[DefaultMPOO] [bit] NOT NULL
) ON [PRIMARY]
GO
CREATE TABLE [MPOO_Zips](
[MPOOID] [int] NOT NULL,
[AreaID] [int] NOT NULL,
[DistrictID] [int] NOT NULL,
[ZipCode] [varchar](5) NOT NULL
) ON [PRIMARY]
GO
insert into MPOO_List values(1000,'PDC','Mary','Jane','[phone]', 'mary.jane#here.com', 5,23,'2017-03-17 15:49:57',0);
insert into MPOO_List values(1001,'NDC','John','Doe','[phone]','john.doe#there.com',5,23,'2015-10-01 11:17:08',0);
insert into MPOO_List values(1103,'MPOO 2','Ann','Nonymous','[phone]','ann.Nonymous#here.com',5,55,'2018-03-23 09:59:06',0);
insert into MPOO_List values(1104,'MPOO 3','Homer','Simpson','[phone]','homer.simpson#fbi.gov',5,55,'2018-03-23 09:55:50',0);
insert into MPOO_List values(2203,'MPOO 2','Bart','Simpson','[phone]','bart.simpson#nasa.gov',11,44,'2016-09-20 06:26:43',0);
insert into MPOO_List values(2204,'MPOO 0','Lisa','Simpson','[phone]','lisa.simpson#mtv.gov',11,44,'2016-09-17 05:45:14',0);
insert into mpoo_zips values (1000,5,23,50265);
insert into mpoo_zips values (1000,5,23,50266);
insert into mpoo_zips values (1000,5,23,50307);
insert into mpoo_zips values (1001,5,23,50533);
insert into mpoo_zips values (1001,5,23,50535);
insert into mpoo_zips values (1001,5,23,50536);
insert into mpoo_zips values (1103,5,55,56341);
insert into mpoo_zips values (1103,5,55,56342);
insert into mpoo_zips values (1103,5,55,56353);
insert into mpoo_zips values (1104,5,55,56450);
insert into mpoo_zips values (1104,5,55,56449);
insert into mpoo_zips values (1104,5,55,56452);
insert into mpoo_zips values (2204,11,4,20006);
insert into mpoo_zips values (2204,11,44,20033);
insert into mpoo_zips values (2204,11,44,20035);
Assuming I want data for AreaID 11, DistrictID 44 to be on top.

Answer 1

You could use UNION to concatenate two queries, the first with the current area and district selection (WHERE AreaID = 5 AND DistrictID = 8), and the second with its opposite (WHERE NOT(AreaID = 5 AND DistrictID = 8).
SELECT /* some columns */ FROM mytable
WHERE AreaID = 5 AND DistrictID = 8
UNION
SELECT /* the same columns */ FROM mytable
WHERE NOT(AreaID = 5 AND DistrictID = 8)
UNIONed queries also seem to be reordered after being united - get around this by selecting an integer (which you will use for your sort order) and giving it a name, and putting all that inside a subquery that orders by the name of your integer.
SELECT * FROM
(
/* UNIONed query goes here */
)
ORDER BY manual_order
There's a catch - you must make sure that the columns you select have unique names. If they don't, the superquery gets confused. You can work around it by aliasing the column names if you need to.
SELECT ml.ID AS mlID, zc.ID as zcID
A user with AreaID of 5 and DistrictID of 0 has edit access to all data within his/her Area which spans multiple districts.
You were right about the problem being in the join, the problem is ml.DistrictID = zc.DistrictID, what should happen if the DistrictID is 0. On reflection, I can't see why you aren't joining on the MPOOID column.
Same thing happens if I add additional columns to SELECT
The UNION result is automatically deduplicated, so if you remove columns that aren't unique, you will find the result set shrinks...
Would I be able to also order by MPOOName within each group?
Just add it as a secondary column in the ORDER BY clause.
ORDER BY manual_order, MPOOName
SQLFiddle complete at http://sqlfiddle.com/#!18/0e442/57
Unrelated, but if you have control of your database design and 5 minutes of spare time, I'd recommend unifying the casing of your table and column names. If your table names are lowercase then use lowercase column names too (and underscores to separate words), else use PascalCase. Everyone's got their preference, for me I prefer the contrast of CAPS for SQL keywords, and lowercase for table/column names.
I'm assuming that your ASP.NET application already knows the AreaID and DistrictID, so you might want to use a builder pattern that checks whether DistrictID > 0, and only include the extra conditions when that returns true.
districtJoin = '';
districtWhere1 = '';
districtWhere2 = '';
if (mlDistrictID > 0)
{
districtJoin = ' AND ml.DistrictID = zc.DistrictID';
districtWhere = ' AND DistrictID = ' + mlDistrictID.ToString();
}
string queryString =
'SELECT /* ml and zc columns */ FROM MPOO_List ml'
+ ' LEFT JOIN ZIPCodes zc ON ml.AreaID = zc.AreaID' + districtJoin
+ ' WHERE ml.AreaID = ' + mlAreaID.ToString() + districtWhere1
+ ' UNION '
+ ' SELECT /* ml and zc columns */ FROM MPOO_List ml'
+ ' LEFT JOIN ZIPCodes zc ON ml.AreaID = zc.AreaID ' + districtJoin
+ ' WHERE NOT(ml.AreaID = ' + mlAreaID.ToString() + districtWhere + ')';