PROS And CONS Of Seperate Databases For CACHING...
May 9, 2006
I have a main database...for this large Web site...and Im wondering
what would be the PROS and CONS of using another database (located on
the same, or on another SQL Server). Im just thinking this would be
good incase we ever needed to take some load off one of the servers.
Also, we will be integrating Community Server into this Web site. Of
course you know CS adds its own database objects which crowd up our
main database objects.
We were thinking of giving CS its own database also; bad practice, or....it doesn't matter much?
My company is thinking about moving to a product that uses Microsoft SQL Server and I have been asked to find out what are the Pros and Cons of the product, if any.
Anyone know where I could find some good articles about pros and cons of upgrading SQL 7 to SQL 2K?
Our boss suddenly got this bug up his rear that we need to upgrade. :( Probably because we are to a point where everything is running smoothly, and he needs to throw a monkey wrench in it.
We are a very small shop, and are not stressing SQL 7 in the least. I am interested to read some articles about SQL 2K, because right now, I can't see any possible reason an upgrade would be worth doing.
I have a requirement where I need to perform a query for positioninformation. But for some types of entries, I need to "expand" the rowto include additional position rows. Let me explain with an example:An index is a security that is made up of components where eachcomponent has a "weight" or a number of shares. So if I have 1 share ofthe index, I have X shares of each component.AAPL is an Equity, CSCO is an Equity, SPY is an Index. Lets say thatSPY has one component, AAPL, with shares being 10. (1 share of SPY = 10shares of AAPL).So, I do some trading and I end up with positions as follows:+10 AAPL-5 CSCO+2 SPYThe query I need returns:+10 AAPL-5 CSCO+2 SPY+20 AAPL (from 2 SPY * 10 shares)which becomes (after grouping):+30 AAPL-5 CSCO+2 SPY-----------------------------------------Based on that criteria and the following schema (and sample data):-- Drop tablesDROP TABLE [SecurityMaster]DROP TABLE [Position]DROP TABLE [IndexComponent]-- Create tablesCREATE TABLE [SecurityMaster] ([Symbol] VARCHAR(10), [SecurityType] VARCHAR(10))CREATE TABLE [Position] ([Account] VARCHAR(10), [Symbol] VARCHAR(10), [Position] INT)CREATE TABLE [IndexComponent] ([IndexSymbol] VARCHAR(10), [ComponentSymbol] VARCHAR(10), [Shares] INT)--Populate tablesINSERT INTO [SecurityMaster] VALUES ('AAPL', 'Equity')INSERT INTO [SecurityMaster] VALUES ('MSFT', 'Equity')INSERT INTO [SecurityMaster] VALUES ('MNTAM', 'Option')INSERT INTO [SecurityMaster] VALUES ('CSCO', 'Equity')INSERT INTO [SecurityMaster] VALUES ('SPY', 'Index')INSERT INTO [Position] VALUES ('001', 'AAPL', 10)INSERT INTO [Position] VALUES ('001', 'MSFT', -5)INSERT INTO [Position] VALUES ('001', 'CSCO', 10)INSERT INTO [Position] VALUES ('001', 'SPY', 15)INSERT INTO [Position] VALUES ('001', 'QQQQ', 21)INSERT INTO [Position] VALUES ('002', 'MNTAM', 10)INSERT INTO [Position] VALUES ('002', 'APPL', 20)INSERT INTO [Position] VALUES ('003', 'SPY', -2)INSERT INTO [IndexComponent] VALUES ('SPY', 'AAPL', 25)INSERT INTO [IndexComponent] VALUES ('SPY', 'CSCO', 50)INSERT INTO [IndexComponent] VALUES ('QQQQ', 'AAPL', 33)-- *****************************-- Based on the rules:-- 1) Index positions appear like other positions (account /symbol) pair, but-- its components show up as new rows of account (of index),symbol (equal--to component symbol), position (equal to shares * index position)-- 2) One row for each account / symbol pair (GROUP BY account andsymbol, SUM position)-- Expected output (without grouping) (sorted by account / symbol)-- 001 AAPL 10-- 001 AAPL 375 (component shares * index position) (25* 15) (SPY)-- 001 AAPL 693 (component shares * index position) (33* 21) (QQQQ)-- 001 CSCO 10-- 001 CSCO 750 (component shares * index position) (50* 15) (SPY)-- 001 MSFT -5-- 001 QQQQ 21-- 001 SPY 15-- 002 AAPL 20-- 002 MNTAM 10-- 003 AAPL -50 (component shares * index position) (25* -2) (SPY)-- 003 CSCO -100 (component shares * index position) (50* -2) (SPY)-- 003 SPY -2-- Expected output (with grouping account / symbol) (sorted by account/ symbol)-- 001 AAPL 1078-- 001 CSCO 760-- 001 MSFT -5-- 001 QQQQ 21-- 001 SPY 15-- 002 AAPL 20-- 002 MNTAM 10-- 003 AAPL -50-- 003 CSCO -100-- 003 SPY -2---------------------------------------------Is a UNION the best way to perform the query. What are the pros andcons? What, if any, is a better way?SELECT[Account], [Symbol], SUM([Position]) AS [Position]FROM(SELECT[Account], [Symbol] , [Position]FROM[Position]UNION ALLSELECTP.[Account] , IC.[ComponentSymbol] AS [Symbol] , (P.[Position] *IC.[Shares]) AS [Position]FROM[IndexComponent] ICJOIN[Position] PONP.[Symbol] = IC.[IndexSymbol]) DGROUP BY[Account], [Symbol]ORDER BY[Account], [Symbol]
Hi alli am building a SQL 2000 database that it is proving a littlechallenging, i have companies with multiple addresses, phone numbers,owning mine sites etc and also joint ventures so maybe you get thepicture with a few design issues that i ma encounteringMy queriy is about a primary key identity, and which one to use withrespect to either the identity data type or the unique identifier ,I am aessentiall building an address table to hold all multipleaddresses as well as phone numbers etc, so my desire to have a uniqueidentity for each record is very important.My view is i will run in to violation errors by just using the tableidentity data type, i could i suppose use composit primary keys butthat may have a performance impact, although thiis will not be a hightransaction database.Does anyone know about performance issues regarding each identitysolution, by using a generated 16 bit identifier there are going to behuge numbers for the DB to verify. or am i worried about nothing?any views greatly appreciatedregardsGreg
Im about to start converting code to Stored Procedures for all my reports in Reporting Services. I was wondering what the pros and cons of this may be.
It seems to me that either by accident or design, SQL Server tends to steer you away from backing up directly over the network.
Are there reasons for not doing this because you obviously don't want to leave your backups on a local drive in case the drive fails?
Some possibilties that I can think of are:-
1. Local drives have faster access times and SQL backups can get quite large. I did a quick test and found that a netwrok backup takes 2 to 3 times longer than it does on a local drive. 2. Backing up on the netwrok could hog too much bandwidth. I haven't tested this and would be surprised it it's true. 3. There could be some reason that you don't want the Server and Agent services running under a domain account but want to leave them on the Local System account. I am not aware of any such reasons by the way. 4. Local drives persumably have a slightly higher availability than network drives. If the server is running, the drive should be available.
Like in the subject: What are the cons and pros of using nvarchar(max) versus ntext? Does it have something to do with having to enable full text search perhaps in the latter case?
I'm just wondering what any pros and cons of using multivalue columns in a database are.
I'm designing a database which will have a column for FABRIC_TYPES_AVAILABLE for a certain FURNITURE_ITEM. Each FURNITURE_ITEM can have multiple FABRIC_TYPES_AVAILABLE of course. So I was just going to store a 2 or 3 digit number of the FABRIC_TYPES_AVAILABLE in that row. So I would have something like....34,24,453,32,23,45,67,65,43,21,21,45.
Anyway....thanks in advance for any information. Links I could read would be great too...b/c I did do a bit of searching, but didnt find much.
I was wanting to know. I am making a site that might be come big. And me and this dude are considering pickering abotu GUID. I don't want to use them but he does. And I was wondering what should we do? I know nothing about guids
vinod writes "Q1>Should I apply filter in sequence(based on primarykey,not null,comparision,between clause) ?
-How to apply filter in correct format,and SQL server internally execute it. Q2>Should I use 'is not Null' to be put at the last?
Q3>Should I use 'between clause' rather than relation operator i.e(empid>10 and empid<200)
Q4>Does filters of sequence has any impact on the query execution process
table1[pkey1,col1,col2,col3] -->pkey is pkey1 table2[pkey2,pkey1,col11,col12,col13]--->pkey is pkey2, and pkey1 is foreign key
CaseI->select table2.pkey2,table1.col1 ,table2.col12 from table1 inner join table2 on table1.pkey1=table2.pkey1 where (table1.col3>100 and table1.col3<300 ) and table2.pkey1=2020 and table2.col13 is not null
CaseII->select table2.pkey2,table1.col1 ,table2.col12 from table1 inner join table2 on table1.pkey1=table2.pkey1 where (table1.col3 between 101 and 299 ) and pkey1=2020 and table2.col13 is not null"
Are there any vices to using default constraints on all columns in your table. For example an Int that defaults to 0 or a char or varchar that defaults to ''
I know that 0 and Null are not the same thing. But if your programs don't have the concept of NULL then you have to convert the NULL to zero.
So, DEFAULT CONSTRAINTS on every column. Is it good or Bad?
Can someone provide information or a link to information regarding the pros and cons of saving files directly into a SQL 2005 database? I'm actually for saving files to a database (cleaner implementation then just saving the location then having to get the file, etc), but my project manager is not convinced so I need to make an argument for (or against depending on what I actually find out) using varbinary data type. Thanks.
Hi All!We are doing new development for SQL Server 2000 and also moving fromSQL 7.0 to SQL Server 2000.What are cons and pros for using IDENTITY property as PK in SQL SERVER2000?Please, share your experience in using IDENTITY as PK .Does SCOPE_IDENTITY makes life easier in SQL 2000?Is there issues with DENTITY property when moving DB from one serverto another? (the same version of SQL Server)Thank you in advance,Andy
hi all, I was wondering if anyone knows of any pros/cons on storing SSIS (2005) packages within SQL Server 2005 SP2. We're contemplating the migration/storage of a large number of packages, a minority of which designed in dts (SQLServer 2000) and using third party activx components.
We are in the process of replacing our primary production server. In the process of determining how SQL server is going to be structured, it has been suggested that I place all current and new indexes on a separate file group. These filegroups would then reside on a separate shelf on the server. What are the pros and cons of doing this?
I am using transaction replication between a transaction and reporting database server. When I use a snapshot to initialize my subscribers, I currently get a lot of deadlocks during the snapshot creation. I am considering using a database backup instead. Can anyone tells me how to reduce the table locks that I am getting during snapshot creation or advice on using database backups?
Hi. I have been talking with some developers who have built a hosted application supporting multiple customers. Their database approach is to create a new, dedicated database (same schema each time) for every customer that signs-up.
This approach is contrary to typical hosted DB designs that I have delt with -- that is, a single database holding multiple customer information rather than a unique database for each customer.
Does the improved security of a dedicated database out-weigh the additional maintenance requirements?
If anyone has some objective thoughts on this topic, I'd love to hear them.
My boss has asked me to look into this and I haven't been able to find any information on the web. I hope someone can answer this for me. We currently have a single database that is storing all the user information and transactions. Within the same database we are also logging different types of user activity. If both these tables are heavily used, would it make sense to separate it into different database, one for data and one for logging? Is there any pro or cons of having more than one database? Any opinion or suggestion would be greatly appreciated. I'm the closest thing they have to DBA and I'm really new to this. Thanks.
We are looking at expanding our web application (C# and Sql Server). The module we are going to be adding will be far more processor intense (plotting and modeling) than any of the current code.
I need to know what the best design for this add on would be.
Would it be better to a) Keep the new module as part of the core db (entire application on 1 database) and use loading balncing between 2 (or more) servers to handle the huge increase in demand? b) Create the new module as a seperate database that runs on a seperate server?
If B is the best option (I am thinking it is not) is there a way to write easy and efficient queries and views of the 2 databases as 1.
Thanks
Justin
Justin Bezanson www.justinbezanson.com www.geekdaily.net - web development and technology blog and news www.offsidegames.com - free online flash games www.thrufare.com - free proxy website
I've got a customer who thinks their data should be sepeated and reside on two seperate databases in case one is compromised. Now I've never heard of anyone doing this and would like comments from the users here as to why this should/(shoult not) be done.
I'm assuming that they want to keep their general info e.g. name and address on one database and keep other (medical info) on the other db.
None of our db's are outside of the firewalls so to me this makes no sense but I would like feedback.
I have two databases in sql server. I'll call them DB1 and DB2. I have a table in DB2 that needs to form a relationship with a table in DB1. When I attempt to add a relationship I only see tables in DB2. Can this be done?
Hello everyone, Here's my situation... I'm running a web service which involves 51 seperate servers and databases. There are fifty licensee servers (One for each US state) and one corporate server. Each night I need to upload sales and membership data from the licensee's databases to the corporate database to compile reports. The application platform I'm using is ASP.NET 2.0 and the the database is SQL2005 express. I want this process to be run automatically, so I believe it's a scheduled windows service I need to setup up in .NET to make the data transfers. If anyone has already set something up like this, or knows the steps to take? I would love to have your input. Thanks in advance, Robert
When this fires a query is done on the customer table to get some values and put some order/customer data into an audit trail table.
In one period of time, it appears that the trigger did not put the data into the separate table. No known reason.
Suspicion: What if the customer table were locked by some process when the trigger fired. Maybe the trigger was chosen as a deadlock-victim.
?? I guess I need to check for @@ERROR during the trigger and do something. Any suggestions? I think I can raise the priority of my trigger to "win" during a deadlock.
If trigger activity is chosen as a deadlock-victim, can the trigger make another attempt to complete it's activity?
I have got MSSQL 2000 set up on a machine in my rack at my local telehouse, and a web server set up at home on an ADSL line.
Both servers can see (ping) eachother fine , so you can rule out any kind of connectivity issues straight away, but when i try to get my forum to connect to the mssql database using the correct credentials it just fails saying that the credentials are incorrect ot the server does not exist.
I also installed an SQL database tool on my web server (Shusheng SQL Tool) and attempted to connect to my SQL server using that tool, and got the following message: '[DBNETLIB][ConnectionOpen (Connect()).]SQL Server does not exist or access denied.'
The server is currently using mixed mode authentication (SQL/Windows) and has both TCP/IP and Named pipes enabled.
Is there some kind of 'Enable remote connections' option in SQL? I need to be able to allow connections to my SQL server from any system, anywhere...
I need to update a table on our Test Server which is GCSQLTEST, with another table thats on our live server GCSQL. How would I go about doing that in a stored procedure??
CREATE PROCEDURE [InsertRevised_MainTable] AS INSERT INTO dbo.RevisedMainTable ([IR Number], [Date], [I/RDocument], [Violation Type]) SELECT [Incident Report No], [Date], [I/RDocument], TypeOfIncident FROM dbo.RevisedMainTable WHERE NOT EXISTS (SELECT * FROM dbo.RevisedMainTable WHERE [IR Number] = [IR Number])
