A new problem with zipcodes..Challenging

I'm not sure I understand your question, but if you want to get the minimum and maximum zips for each code, this will do it:

SELECT zmin, zmax, a.code FROM
(SELECT Code, max(zip) as zmax FROM TableA GROUP BY Code ) as a INNER JOIN
(SELECT Code, min(zip) as zmin FROM TableA GROUP BY Code) b ON a.code = b.code

actualy the data ove needs to be changed..
zip                            code
     381370000          Q328
     381383904          Q104
     381383929          Q104
     381383930          Q104
     381389999          Q328
     381390000          Q328

The range shud be
zmin                  zmax           code
38170000        381383903     Q238
381383904      381383904     Q104
381383905      381383928     Q328
381043929      381383930     Q104
381043931      381389999      Q328

the table has again more then 100000 records......and performance is of course an issue....
so if the base exists than we shud have a range for the base and the base then needs to be split in to plus four

i.e if i have just this
381051235    Q111
381059999    Q123
381069999    Q123
theh the range for this will be

zmin               zmax           code
381050000   381051234       Q123
381051235   381051235       Q111
381051236   381069999       Q133

the numbers will always be arranged in order ....
thanks Guys...

Look at these posts these will help understanding that the problem is not that simple......
there are around 500 codes and 1000000 zipcodes..
Well the problem is that the range will split....the min and max will then result in overlaps.....

https://www.experts-exchange.com/questions/21076168/Sql-Server-Query.html

This can be done easily using DTS.  I noticed people in your previous questions used queries and temp tables but in my opinion this is too complicated, although it might provide somewhat better performance.  If you're interested in a solution using DTS let me know and I will elaborate.

Sure DTS will help too...any solutions will definitely help me resolve the problem, Please let me what do u have in yur mind.

Thanks.

Yes, we're all ears.

I ran it for sometime but it was taking too much time to run..again as i have 100000 records it takes too much time...i will run it in steps and ssee if it gives right results

What kind of hardware are you running this on?

SQL Server 200o and i think i have a P4

What kind of hard drives? You might want to consider a low-level Xeon/Opteron system with 2 or 3 drives (preferably SCSI) in a RAID 0 or 5 configuration.

Also, how much memory? What is the size of the table (in KB)?

Furthermore, consider creating indexes on VOwner and VNotOwner (on ZIP); that could speed things up quite nicely.

Just dropping in to say I see you are still at it Rick.   I'd like to help and Sjoerd is one of the most capable experts here, but I think your reoccuring problem is that you are not explaining the problem in a way we can understand.

I wonder if you could explain where the data comes from, what you want to do with it, and why?  That's usually the opposite of what we usually ask - just give us the data input and desired output is the norm - but we might be trying to develop a solution around your preconceptions of how it needs to be done when there is another more efficient solution.

Sorry got back so late in here guys Well let me explain the problem amd may be u guys can work with the logic i was trying to use the logic i thought would work best to get the ranges...

I have a database which has 50000 zipcodes, these zipcodes are assigned to codes as i had suggested..
The way the data is in the table is that
i have a cloumn which has only 5 digits and another column which has 4 digit zips....
the primary key is a combination of these two columns
eg
91234 9999

The way codes are assigned to the zipcodes is that if the base owner i.e the owner of 91234 9999 will get all the zip codes in the zip 91234..i.e 91234 0000 to 91234 9999 except for if someone else is specifically assigend a code in that +4 i.e

91234 9999 A101 means everything in 91234 0000 to 91234 9999 should go to A101 but we do not have the +4's in the database becasue there are no exceptions to this so i will just have 91234 9999
but i can have exceptions like i have a case where
91233 9999 A100
91234 9999 A123
91235 9999 A101
91235 1234 A123
91235 2000 A123
91236 9999 A101

if we look above the range should be
91233 0000    91233 9999 A100
91234 0000    91234 9999 A123
91235 0000    91235 1233 A101
91235 1234    91235 1234 A123
91235 1235    91235 1999 A101
91235 2000    91235 2000 A123
91235 2001    91236 9999 A101

thats just one case
lets look athe the other

91233 9999 A123
91234 9999 A123
91235 9999 A101
91235 1234 A123
91235 2000 A123
91236 9999 A123

Now the ranges would change

91233 0000   91234 9999 A123
91235 0000   91235 1233 A101 < because the base owner i.e owner of 9999 is A101
91235 1234   91235 1234 A123
91235 1235   91235 1999 A101
91235 2000   91235 2000 A123
91235 2001   91235 9999 A101
91236 0000   91236 9999 A123

now lets look at another example

91789 9999   A101
91790 9999   A101
91799 9999   A101 <-----the gap here should be identifed as a gap..and then assigned to the right code which will be A101
91801 9999   A101
91803 9999   A101
91804 9999   A101
91805 9999   A102
91806 9999   A102
91807 9999   A102
91808 9999   A102

now here my ranges are
91789 0000 91790 9999   A101
91799 0000 91799 9999   A101
91801 0000 91801 9999   A101
91803 0000 91804 9999   A101
91805 0000 91808 9999   A102
the zip codes that are no in here should be assigned using the gaps logic becasue they are actualy gaps......if we can get the ranges i can get the gaps from there but the tough part is to get the ranges and handle all these exceptions....
now just two zipcodes belong to another code rest all are A101..
I need to find the ranges for all the zipcodes as i have suggested...
in the above cases i will have the range broken down...
 I dont have 0000 in +4 ....... only if there is an exception it will have a +4 and thats is why i am not able to find the ranges....
Does this help Kselvia ..
I can go in to more detail if you want me to ..please let me know..thanks s ton for your help guys
   

Have you tried putting indexes on vowner and vnotowner and running the query as I gave it?

Yes i have but it is still slow and takes a long time to run.

In that case, you might just have to use a beefier machine (first, try more memory).