Article View: borland.public.cppbuilder.database.ado
Article #1650Re: Help with recursive queries... BIG problem.
From: mymn
Date: Mon, 08 Mar 2004 02:01
Date: Mon, 08 Mar 2004 02:01
170 lines
6044 bytes
6044 bytes
just my suggestion :- you need to modify your database design to ensure your data accessing will be efficient. the entity relationship between PEOPLE and VEHICLE is ONE-TO-MANY or maybe MANY-TO-MANY. explaination; one PEOPLE can have many VEHICLE. normally, one VEHICLE can only be own by one PEOPLE. but, on your condition, if one PEOPLE related on another PEOPLE and as i'm understood, the VEHICLE related on that another PEOPLE will be own by that one PEOPLE too. my suggestion is; you make an inter-section table which combining the code from VEHICLE and PEOPLE. you can called it PEOPLE_VEHICLE. in table VEHICLE (Code,Make,Color,...) in table PEOPLE(Code,Name,Age,...) in table PEOPLE_VEHICLE(CodePeople,CodeVehicle,Status) in field Status at table PEOPLE_VEHICLE; you can set option if the connection between PEOPLE and VEHICLE, then set PV. but, if the connection between PEOPLE and PEOPLE, then set PP. for example, PEOPLE Code Name Age 132 Alex 30 140 George 27 VEHICLE Code Make Color 100 US Blue 242 US Black 500 US Black PEOPLE_VEHICLE CodePeople CodeVehicle Status 132 100 PV 132 242 PV 132 140 PP 140 500 PV if you do this method, you can only using this inter-section table to access/link on each record related between PEOPLE and VEHICLE table. than, you just query such as like this; select * from PEOPLE_VEHICLE where (CodePeople='132' and Status='PV') or (CodeVehicle='132' and Status='PP') that's all. hope it'll help you. from my experience, the technique of search depend on your database design. if you make normalization for your database design, you'll get the best entity-relationship for your tables. "Star" <lem@lemforeverREMOVE_THIS.com> wrote in message news:40439ba3$1@newsgroups.borland.com... > Hi > > This is a problem that I'm having and I was wondering that maybe you guys > could come up with some ideas. > Ok, let's see if I can explain the problem. > > I have about 15 tables. Each table has several fields, however each table > has a common field called 'Code' > > ex. > > Table Vehicles > -------------- > Code > Make > Color > > Table People > ------------- > Code > Name > Age > > Table Addresses > ----------------- > Code > City > ZIP > ... > > Now we want to relate records from one table to another. For example, we > want to say 'John has two cars' > In order to do that, I have an external table called Links with source an > destination fields. For the example we would have > > Source Dest > ---------------- > P132 V100 > P132 V242 > > (P132 is a record from the People table with code 132, V100 a vehicle from > the Vehicle table with code 100, and V242 a vehicle > with code 242) > > So, with all that, we know that John (P132) is related with two vehicles. If > I want to run the query 'People associated with vehicles' > it would be easy to query that table and get that information. > > Now, let's suppose we have this situation: > > Source Dest > ---------------- > P132 V100 > P132 V242 > P140 V500 > P132 P140 > > In this example, John is related with Michael (P140) and Michael is related > with a vehicle (V500). > If I want to run the same query than before 'People associated with > vehicles', because of P132 is related with P140 and P140 is related with > a V500 this implies that P132 is related with V500. > > Now, the query is not so simple, because it's like a graph and I need to > make a Depth First Search to get what I want. Doing that, I will get that > P132 is > related with V500 because there is a 'path' to get there. > > I have done all this and it works fine. The problem that I have is that I > need to make that Depth First Search for each people that I have in the > database > in order to see if I have vehicles associated with it. If I don't have many > records it's ok, but If I have 100,000 people on the database I would have > to > make that search 100,000 times just to see if there is a path between them, > I mean, if there is a way to get from a person to a vehicle. > In this case... this solution doesn't work, because it takes forever... > > > What do you guys suggest I can try? I though that one solution would be to > have a process running on the server that maintains a table with ALL the > links that we have in the database and the depth of the link. So, in our > example we'd have a entry "P132 V500 with depth 2" > Having that table it would be easy to query. Inconvenience: the table is > going to be HUGE, and also the process has to be smart enough to keep track > of > deleted links, new links... > > I'd appreciate if you guys have any idea or suggestions about this and how I > can get a better perfomance. > > Thanks a lot for reading this whole message!! > > > > > > >
Technical Details
Message-ID:
<404bd3f4@newsgroups.borland.com>Path:
arc1-ca.pugleaf.net!archiver1.google.com!news1.google.com!news2.google.com!fu-berlin.de!newsfeed01.sul.t-online.de!t-online.de!newsgroups.borland.com!not-for-mailReferences:
<40439ba3$1@newsgroups.borland.com>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 10698b,31dd6211c454d261 X-Google-Attributes: gid10698b,public X-Google-ArrivalTime: 2004-03-07 18:09:50 PST From: "mymn" <yusairi@yahoo.com> Newsgroups: borland.public.cppbuilder.database.ado References: <40439ba3$1@newsgroups.borland.com> Subject: Re: Help with recursive queries... BIG problem. Date: Mon, 8 Mar 2004 10:01:38 +0800 Lines: 172 X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2720.3000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2727.1300 NNTP-Posting-Host: 219.93.61.50 Message-ID: <404bd3f4@newsgroups.borland.com> X-Trace: newsgroups.borland.com 1078711284 219.93.61.50 (7 Mar 2004 18:01:24 -0800) Path: arc1-ca.pugleaf.net!archiver1.google.com!news1.google.com!news2.google.com!fu-berlin.de!newsfeed01.sul.t-online.de!t-online.de!newsgroups.borland.com!not-for-mail