GUIDs as Primary Keys - Offline OLTP - sql-server

We are working on designing an application that is typically OLTP (think: purchasing system). However, this one in particular has the need that some users will be offline, so they need to be able to download the DB to their machine, work on it, and then sync back once they're on the LAN.
I would like to note that I know this has been done before, I just don't have experience with this particular model.
One idea I thought about was using GUIDs as table keys. So for example, a Purchase Order would not have a number (auto-numeric) but a GUID instead, so that every offline client can generate those, and I don't have clashes when I connect back to the DB.
Is this a bad idea for some reason?
Will access to these tables through the GUID key be slow?
Have you had experience with these type of systems? How have you solved this problem?
Thanks!
Daniel

Using Guids as primary keys is acceptable and is considered a fairly standard practice for the same reasons that you are considering them. They can be overused which can make things a bit tedious to debug and manage, so try to keep them out of code tables and other reference data if at all possible.
The thing that you have to concern yourself with is the human readable identifier. Guids cannot be exchanged by people - can you imagine trying to confirm your order number over the phone if it is a guid? So in an offline scenario you may still have to generate something - like a publisher (workstation/user) id and some sequence number, so the order number may be 123-5678 -.
However this may not satisfy business requirements of having a sequential number. In fact regulatory requirements can be and influence - some regulations (SOX maybe) require that invoice numbers are sequential. In such cases it may be neccessary to generate a sort of proforma number which is fixed up later when the systems synchronise. You may land up with tables having OrderId (Guid), OrderNo (int), ProformaOrderNo (varchar) - some complexity may creep in.
At least having guids as primary keys means that you don't have to do a whole lot of cascading updates when the sync does eventually happen - you simply update the human readable number.

#SqlMenace
There are other problems with GUIDs, you see GUIDs are not sequential, so inserts will be scattered all over the place, this causes page splits and index fragmentation
Not true. Primary key != clustered index.
If the clustered index is another column ("inserted_on" springs to mind) then the inserts will be sequential and no page splits or excessive fragmentation will occur.

This is a perfectly good use of GUIDs. The only draw backs would be a slight complexity in working with GUIDs over INTs and the slight size difference (16 bytes vs 4 bytes).
I don't think either of those are a big deal.

Will access to these tables through
the GUID key be slow?
There are other problems with GUIDs, you see GUIDs are not sequential, so inserts will be scattered all over the place, this causes page splits and index fragmentation
In SQL Server 2005 MS introduced NEWSEQUENTIALID() to fix this, the only problem for you might be that you can only use NEWSEQUENTIALID as a default value in a table

You're correct that this is an old problem, and it has two canonical solutions:
Use unique identifiers as the primary key. Note that if you're concerned about readability you can roll your own unique identifier instead of using a GUID. A unique identifier will use information about the date and the machine to generate a unique value.
Use a composite key of 'Actor' + identifier. Every user gets a numeric actor ID, and the keys of newly inserted rows use the actor ID as well as the next available identifier. So if two actors both insert a new row with ID "100", the primary key constraint will not be violated.
Personally, I prefer the first approach, as I think composite keys are really tedious as foreign keys. I think the human readability complaint is overstated -- end-users shouldn't have to know anything about your keys, anyways!

Make sure to utilize guid.comb - takes care of the indexing stuff. If you are dealing with performance issues after that then you will be, in short order, an expert on scaling.
Another reason to use GUIDs is to enable database refactoring. Say you decide to apply polymorphism or inheritance or whatever to your Customers entity. You now want Customers and Employees to derive from Person and have them share a table. Having really unique identifiers makes data migration simple. There are no sequences or integer identity fields to fight with.

I'm just going to point you to What are the performance improvement of Sequential Guid over standard Guid?, which covers the GUID talk.
For human readability, consider assigning machine IDs and then using sequential numbers from those machines as a possibility. This will require managing the assignment of machine IDs, though. Could be done in one or two columns.
I'm personally fond of the SGUID answer, though.

Guids will certainly be slower (and use more memory) than standard integer keys, but whether or not that is an issue will depend on the type of load your system will see. Depending on your backend DB there may be issues with indexing guid fields.
Using guids simplifies a whole class of problems, but you pay for it part will performance and also debuggability - typing guids into those test queries will get old real fast!

The backend will be SQL Server 2005
Frontend / Application Logic will be .Net
Besides GUIDs, can you think of other ways to resolve the "merge" that happens when the offline computer syncs the new data back into the central database?
I mean, if the keys are INTs, i'll have to renumber everything when importing basically. GUIDs will spare me of that.

Using GUIDs saved us a lot of work when we had to merge two databases into one.

If your database is small enough to download to a laptop and work with it offline, you probably don't need to worry too much about the performance differences between ints and Guids. But do not underestimate how useful ints are when developing and troubleshooting a system! You will probably need to come up with some fairly complex import/synch logic regardless of whether or not you are using Guids, so they might not help as much as you think.

#Simon,
You raise very good points. I was already thinking about the "temporary" "human-readable" numbers i'd generate while offline, that i'd recreate on sync. But i wanted to avoid doing with with foreign keys, etc.

i would start to look at SQL Server Compact Edition for this! It helps with all of your issues.
Data Storage Architecture with SQL Server 2005 Compact Edition
It specifically designed for
Field force applications (FFAs). FFAs
usually share one or more of the
following attributes
They allow the user to perform their
job functions while disconnected from
the back-end network—on-site at a
client location, on the road, in an
airport, or from home.
FFAs are usually designed for
occasional connectivity, meaning that
when users are running the client
application, they do not need to have
a network connection of any kind. FFAs
often involve multiple clients that
can concurrently access and use data
from the back-end database, both in a
connected and disconnected mode.
FFAs must be able to replicate data
from the back-end database to the
client databases for offline support.
They also need to be able to replicate
modified, added, or deleted data
records from the client to the server
when the application is able to
connect to the network

First thought that comes to mind: Hasn't MS designed the DataSet and DataAdapter model to support scenarios like this?
I believe I read that MS changed their ADO recordset model to the current DataSet model so it works great offline too. And there's also this Sync Services for ADO.NET
I believe I have seen code that utilizes the DataSet model which also uses foreign keys and they still sync perfectly when using the DataAdapter. Havn't try out the Sync Services though but I think you might be able to benefit from that too.
Hope this helps.

#Portman By default PK == Clustered Index, creating a primary key constraint will automatically create a clustered index, you need to specify non clustered if you don't want it clustered.

Related

Which PK type to use Int or BigInt or UniqueIdentifier

Our team developing a new database for health care ERP. During the brain storming meeting I recommended to use the uniqueidentifier because it has many benefits like
Less round trip to the database OnInsert if we generate the value from client application
By generating it on the client application, we can use more easily the master-detail approach.
It helps in data replication
Till now, I was confident and even I thought I would hear some compliments, till my boss asked me couple of questions:
You are going to use this Guid as primary key with clustered indexing? .
Do you know the size of your table how big it and its consequences on the performance?
Some of the developers proposed the Int and others BigInt
I would like to know if my Boss questions have a base or what I am thinking is true because what I think is best thing for building ERP with replication support.
NOTE I did already search for long time here in this site and on other sites also.
Which of the above is the best key to be used in ERP like health care information system?
Think about what your company is proposing to do and the level of expertise your group currently has. Apparently it does not have significant experience with sql server based on your questions and your manager's questions. I cannot reasonably see a way for you to develop an enterprise-scale system without the necessary expertise - especially with the backend systems that you plan on using.
And your process (as little as you describe it) sounds concerning. "Brainstorming" is not, IMO, a point where you decide on schemas and choose keys. And one should not just blindly choose a particular datatype for every primary key. But all of this is guessing without knowing more about where you are in this process. If your schema is not yet fixed (regardless of what datatypes are selected for each column), then you are not yet in a position to worry about performance.
Lastly, you and your manager confuse two related but independent attributes. A primary key is not the same as the clustered index, despite the unfortunate implementation choices made by the MS development team. They are independent of each other; make a conscious decision about your clustered indexes and do not allow the db engine to automatically choose the primary key as the clustered index.
So to answer your questions. Yes - those questions are valid. But your project does not yet appear to have reached a point where those concerns can be addressed.

Database Designing: An art or headache (Managing relationships)

I have seen in my past experience that most of the people don't use physical relationships in tables and they try to remember them and apply them through coding only.
Here 'Physical Relationships' refer to Primary Key, Foreign Key, Check constraints, etc.
While designing a database, people try to normalize the database on paper and keep things documented. Like, if I have to create a database for a marketing company, I will try to understand its requirements.
For example, what fields are mandatory, what fields will contain only (a or b or c) etc.
When all the things are clear, then why are most of the people afraid of the constraints?
Don't they want to manage things?
Do they have a lack of knowledge
(which I don't think is so)?
Are they not confident about future
problems?
Is it really a tough job managing all these entities?
What is the reason in your opinion?
I always have the DBMS enforce both primary key and foreign key constraints; I often add check constraints too. As far as I am concerned, the data is too important to run the risk of inaccurate data being stored.
If you think of the database as a series of stored true logical propositions, you will see that if the database contains a false proposition - an error - then you can argue to any conclusion you want. Given a false premise, any conclusion is true.
Why don't other people use PK and FK constraints, etc?
Some are unaware of their importance (so lack of knowledge is definitely a factor, even a major factor). Others are scared that they will cost too much in performance, forgetting that one error that has to be fixed may easily use up all the time saved by not having the DBMS do the checking for you. I take the view that if the current DBMS can't handle them well, it might be (probably is) time to change DBMS.
Many developers will check the constraints in code above the database before they actually go to perform an operation. Sometimes, this is driven by user experience considerations (we don't want to present choices / options to users that can't be saved to the database). In other cases, it may be driven by the pain associated with executing a statement, determining why it failed, and then taking corrective action. Most people would consider code more maintainable if it did the check upfront, along with other business logic that might be at play, rather than taking corrective action through an exception handler. (Not that this is necessarily an ideal line of thinking, but it is a prevalent one.) In any case, if you are doing the check in advance of issuing the statement, and not particularly conscious of the fact that the database might get touched by applications / users who are not coming in through your integrity-enforcing code, then you might conclude that database constraints are unnecessary, especially with the performance hit that could be incurred from their use. Also, if you are checking integrity in the application code above the database, one might consider it a violation of DRY (Don't Repeat Yourself) to implement logically equivalent checks in the database itself. The two manifestations of integrity rules (those in database constraints and those in application code above the database) could in principle become out-of-sync if not managed carefully.
Also, I would not discount option 2, that many developers don't know much about database constraints, too readily.
Well, I mean, everyone is entitled to their own opinion and development strategy I suppose, but in my humble opinion these people are almost certainly wrong :)
The reason, however, someone may wish to avoid constraints is efficiency. Not because constraints are slow, but because storing redundant data (i.e. caching) is a very effective way of speeding up (well, avoiding) an expensive calculation. This is an acceptable approach, when implemented properly (i.e. the cache is updated a regular/appropriate intervals, generally I do this with a trigger).
As to the motivation to not us FKs without a caching motivation, I can't imagine it. Perhaps they aim to be 'flexible' in their DB structure. If so, fine, but then don't use a relational DB, because it's pointless. Non-relational DBs (OO dbs) certainly have their place, and may even arguably be better (quite arguable, but interesting to argue) but it's a mistake to use a relational DB and not use it's core properties.
I would always define PK and FK constraints. especially when using an ORM. it really makes the life easy for everybody to let the ORM reverse engineer the database instead of manually configuring it to use some PKs and FKs
There are several reasons for not enforcing relationships in descending order of importance:
People-friendly error handling.
Your program should check constraints and send an intelligible message to the user. For some reason normal people dont like "SQL exception code -100013 goble rule violated for table gook'.
Operational flexibility.
You dont really want your operators trying to figure out which order you must load your tables in at 3 a.m., nor do you want your testers pulling their hair out 'cause they cannot reset the database back to its starting position.
Efficiency.
Cheking constraints does consume IO and CPU.
Functionality.
Its a cheap way to save details for later recovery. For instance in an on line order system you could leave the detail item rows in the table when the users kills a parent order, if he later reinstates the order the details re-appear as if by a miracle -- you acheive this extra feature by deleteing lines of code. (course you need some housekeeping process but it is trivial!)
As things get more complex and more tables and relationships are needed in the database, how can you ensure the database developer remembers to check all of them? When you makea change to the schema that adds a new "informal" relationship, how can you ensure all the application code which might be affected gets changed?
Suddenly you could be deleting records that should stay because they have related data the developer forgot to check when writng the delete process or because that process was in place before the last ten related tables were added to the schema.
It is foolhardy in the extreme to not formally set up PK/FK relationships. I process data received from many different vendors and databases. You can tell which ones have data integrity problems most likely caused by a failure to explicitly define relationships by the poor quality of their data.

Few database design questions relating to user content site

Designing a user content website (kind of similar to yelp but for a different market and with photo sharing) and had few databse questions:
Does each user get their own set of
tables or are we storing multiple
user data into common tables? Since
this even a social network, when
user sizes grows for scalability
databases are usually partitioned
off. Different sets of users are
sent separately, so what is the best
approach? I guess some data like
user accounts can be in common
tables but wall posts, photos etc
each user will get their own table?
If so, then if we have 10 million
users then that means 10 million x
what ever number of tables per user?
This is currently being designed in
MySQL
How does the user tables know what
to create each time a user joins the
site? I am assuming there may be a
system table template from which it
is pulling in the fields?
In addition to the above question,
if tomorrow we modify tables,
add/remove features, to roll the
changes down to all the live user
accounts/tables - I know from a page
point of view we have the master
template, but for the database, how
will the user tables be updated? Is
that something we manually do or the
table will keep checking like every
24 hrs with the system tables for
updates to its structure?
If the above is all true, that means we are maintaining 1 master set of tables with system default values, then each user get the same value copied to their tables? Some fields like say Maximum failed login attempts before system locks account. One we have a system default of 5 login attempts within 30 minutes. But I want to allow users also to specify their own number to customize their won security, so that means they can overwrite the system default in their own table?
Thanks.
Users should not get their own set of tables. It will most likely not perform as well as one table (properly indexed), and schema changes will have to be deployed to all user tables.
You could have default values specified on the table for things that are optional.
With difficulty. With one set of tables it will be a lot easier, and probably faster.
That sort of data should be stored in a User Preferences table that stores all preferences for all users. Again, don't duplicate the schema for all users.
Generally the idea of creating separate tables for each entity (in this case users) is not a good idea. If each table is separate querying may be cumbersome.
If your table is large you should optimize the table with indexes. If it gets very large, you also may want to look into partitioning tables.
This allows you to see the table as 1 object, though it is logically split up - the DBMS handles most of the work and presents you with 1 object. This way you SELECT, INSERT, UPDATE, ALTER etc as normal, and the DB figures out which partition the SQL refers to and performs the command.
Not splitting up the tables by users, instead using indexes and partitions, would deal with scalability while maintaining performance. if you don't split up the tables manually, this also makes that points 2, 3, and 4 moot.
Here's a link to partitioning tables (SQL Server-specific):
http://databases.about.com/od/sqlserver/a/partitioning.htm
It doesn't make any kind of sense to me to create a set of tables for each user. If you have a common set of tables for all users then I think that avoids all the issues you are asking about.
It sounds like you need to locate a primer on relational database design basics. Regardless of the type of application you are designing, you should start there. Learn how joins work, indices, primary and foreign keys, and so on. Learn about basic database normalization.
It's not customary to create new tables on-the-fly in an application; it's usually unnecessary in a properly designed schema. Usually schema changes are done at deployment time. The only time "users" get their own tables is an artifact of a provisioning decision, wherein each "user" is effectively a tenant in a walled-off garden; this only makes sense if each "user" (more likely, a company or organization) never needs access to anything that other users in the system have stored.
There are mechanisms for dealing with loosely structured types of information in databases, but if you find yourself reaching for this often (the most common method is called Entity-Attribute-Value), your problem is either not quite correctly modeled, or you may not actually need a relational database, in which case it might be better off with a document-oriented database like CouchDB/MongoDB.
Adding, based on your updated comments/notes:
Your concerns about the number of records in a particular table are most likely premature. Get something working first. Most modern DBMSes, including newer versions of MySql, support mechanisms beyond indices and clustered indices that can help deal with large numbers of records. To wit, in MS Sql Server you can create a partition function on fields on a table; MySql 5.1+ has a few similar partitioning options based on hash functions, ranges, or other mechanisms. Follow well-established conventions for database design modeling your domain as sensibly as possible, then adjust when you run into problems. First adjust using the tools available within your choice of database, then consider more drastic measures only when you can prove they are needed. There are other kinds of denormalization that are more likely to make sense before you would even want to consider having something as unidiomatic to database systems as a "table per user" model; even if I were to look at that route, I'd probably consider something like materialized views first.
I agree with the comments above that say that a table per user is a bad idea. Also, while it's a good idea to have strategies in mind now for how you can cope when things get really big, I'd concentrate on getting things right for a small number of users first - if no-one wants to / is able to use your service, then unfortunately you won't be faced with the problem of lots of users.
A common approach among very large sites is database sharding. The summary is: you have N instances of your database in parallel (on separate machines), and each holds 1/N of the total data. There's some shared way of knowing which instance holds a given bit of data. To access some data you have 2 steps, rather than the 1 you might expect:
Work out which shard holds the data
Go to that shard for the data
There are problems with this, such as: you set up e.g. 8 shards and they all fill up, so you want to share the data over e.g. 20 shards -> migrating data between shards.

Pros and cons of programmatically enforcing foreign key than in database

It is causing so much trouble in terms of development just by letting database enforcing foreign key. Especially during unit test I can’t drop table due to foreign key constrains, I need to create table in such an order that foreign key constrain warning won’t get triggered. In reality I don’t see too much point of letting database enforcing the foreign key constrains. If the application has been properly designed there should not be any manual database manipulation other than select queries. I just want to make sure that I am not digging myself into a hole by not having foreign key constrains in database and leaving it solely to the application’s responsibility. Am I missing anything?
P.S. my real unit tests (not those that use mocking) will drop existing tables if the structure of underlying domain object has been modified.
In my experience, if you don't enforce foreign keys in a database, then eventually (assuming the database is relatively large and heavily used) you will end up with orphaned records. This can happen in many ways, but it always seems to happen.
If you index properly, there should not be any performance advantages to foreign keys.
So the question is, does the potential damage/hassle/support cost/financial cost of having orphaned records in your database outweigh the development and testing hassle?
In my experience, for business applications I always use foreign keys. It should just be a one-time setup cost to get your build scripts working correctly, and the data stability will more than pay for that over the life of an application.
The point of enforcing the rules in the database is that it's declarative - e.g. you do not have to write ton of code to handle it.
As far as your unit tests, just delete tables in the proper order. You just have to write a function to do it right once.
Your issues in development should not drive the DB design. Constantly rebuilding a DB is a developer use case, not a customer use case.
Also, the DB constraints help beyond your application. You never know what your customer might try to do. Don't over do it, but you need a few.
It might seem like you can rely on your applications to follow implied rules, but unless you enforce them eventually someone will make a mistake.
Or maybe 5 years from now someone will do a tidy-up of old records "which are no longer needed" and not realise that there is data in other tables still referencing them. Then a few days/weeks later you or your successor gets the fun job of trying to repair the mess that the database has got in to. :-)
Here's a nice discussion on that in a previous question on SO: What's wrong with foreign keys?. [Edit]: The argument is to make non-enforced foreign keys to get some of the pros if any of the cons apply.
If the application has been properly
designed there should not be any
manual database manipulation other
than select queries
What? What kind of koolaid are you drinking? Most databases applications exist to manipulate the data in the database not just to see it. Generally the whole purpose of the application is to add the new orders or create the new customer records or document the customer service calls etc.
Foreign keys are for data integrity. Data integrity is critical to being able to use the data with any reliability. Databases without data integrity are useless and can cause companies to lose money. This trumps your self-centered view that FKs aren't needed because they make development more complicated for you. The data is far more important than your convenience in running tests (which can be written to account for the FKs).

Loosely Coupled Database Design - How To?

I'm implementing a web - based application using silverlight with an SQL Server DB on the back end for all the data that the application will display. I want to ensure that the application can be easily scalable and I feel the direction to go in with this is to make the database loosely coupled and not to tie everything up with foreign keys. I've tried searching for some examples but to no avail.
Does anyone have any information or good starting points/samples/examples to help me get off the ground with this?
Help greatly appreciated.
Kind regards,
I think you're mixing up your terminology a bit. "Loosely coupled" refers to the desirability of having software components that aren't so dependent upon each other that they can't function or even compile without being together in the same program. I've never seen the term used to describe the relationships between tables in the same database.
I think if you search on the terms "normalization" and "denormalization" you'll get better results.
Unless you're doing massive amounts of inserts at a time, like with a data warehouse, use foreign keys. Normalization scales like crazy, and you should take advantage of that. Foreign keys are fast, and the constraint really only holds you back if you're inserting millions upon millions of records at a time.
Make sure that you're using integer keys that have a clustered index on them. This should make joining table very rapid. The issues you can get yourself wrapped around without foreign keys are many and frustrating. I just spent all weekend doing so, and we made a conscious choice to not have foreign keys (we have terabytes of data, though).
Before you even think of such a thing, you need to think about data integrity. Foreign keys exist so that you cannot put records into tables if the primary data they are based on is not there. If you do not use foreign keys, you will sooner or later (probably sooner) end up with worthless data because you don't really know who the customer is that the order is attached to for instance. Foreign keys are data protection, you should never consider not using them.
And even though you think all your data will come from your application, in real life, this is simply not true. Data gets in from multiple applications, from imports of large amounts of data, from the query window (think about when someone decides to update all the prices they aren't going to do that one price at a time from the user interface). Data can get into database from many sources and must be protected at the database level. To do less is to put your entire application and data at risk.
Intersting comment about database security when data is input through external sources like database scripts.

Resources