The co-founder of Reddit gave a presentation on issues they had while scaling to millions of users. A summary is available here.
What surprised me is point 3:
Instead, they keep a Thing Table and a Data Table. Everything in Reddit is a Thing: users, links, comments, subreddits, awards, etc. Things keep common attribute like up/down votes, a type, and creation date. The Data table has three columns: thing id, key, value. There’s a row for every attribute. There’s a row for title, url, author, spam votes, etc. When they add new features they didn’t have to worry about the database anymore. They didn’t have to add new tables for new things or worry about upgrades.
This seems like a terrible idea to me, but it seems to have worked out for Reddit. Is it a good idea in general, though? Or is it a peculiarity of Reddit that happened to work out for them?
This is a data model known as EAV for entity-attribute-value. It has its uses. A prime example is patient test data which is naturally sparse since there are hundreds of thousands of tests which might be run, but typically only a handful are present for a patient. A table with hundreds of thousands of columns is silly, but a table with EAV makes good sense.
Most of the really big web sites end up using some sort of incredibly simple on the database side of things. This has the advantage that it's fast and scalable. It has the disadvantage that all the relationships that you'd get the database to enforce automatically (via triggers and such) you need to enforce yourself in your client code instead. Maintaining consistency is a pain in the neck, and there's almost always at least some chance that your data will be inconsistent, at least for short periods of time.
For a social networking site, it's a worthwhile compromise. Data that's mostly right most of the time is adequate (e.g., who really cares if the number of up-votes you receive for an item is really 20 milliseconds out of date when it's sent), and keeping costs reasonable while scaling to support a gazillion users matters a lot.
I noted that they did not mention anything about the ease or difficulty in creating reports against that data. When used in a narrow set of circumstances, EAVs can be beneficial. As a central part of most systems it will become a nightmare when you hit reporting. The problem with EAVs is that most of the benefit is at the outset of the project and most of the pain is later in the analysis and reporting especially due to the severe lack of data integrity. "Not having to worry about foreign keys" to me sounds like a nightmare of orphaned rows. Add in the use of surrogate keys for everything and you have a tangled morass which generally ends in a complete rewrite
We worked in a similar problem not long ago, i could say at first it wasn't easy and fun, but after some point that you get used to it, it has it's own benefit, it's like developing another Database withing your tables, in some area it's an overkill task, but when you pass these levels, it provide you with lots of functionality, basically after one point, we didn't create any new table, and we just created dynamic forms for everything's, even for our own programming tasks.
as for performance, system didn't get million of rows to be fair comparison, but for daily usage i never noticed any differences.
some problems i want to share.
we didn't delete any rows we just hide them and setting a flag, and a nightly (weekly) service clean the physical rows
orphan rows, we basically didn't care about cleaning childs, we just set "IsDeleted" On father, and nightly service would clean every rows that is orphan or not needed anymore.
3.you should keep your indexes up to date, but you should skip of building them whenever possible (again nightly service keep index up to date)
we prepared our reporting data ahead of time (AOT) which means we were behind of actual data here :))
we tried hard no to jump in ocean of rows to calc some values by user demand. if we prepared it you can use it, if not then you cant
at the end, there are so many unique challenges to this approach that you should find a way to solve, problems you never faced before in your routine job, but after all of these you earn more flexibility that you can spend.
As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance.
Closed 9 years ago.
Whether we like it or not, many if not most of us developers either regularly work with databases or may have to work with one someday. And considering the amount of misuse and abuse in the wild, and the volume of database-related questions that come up every day, it's fair to say that there are certain concepts that developers should know - even if they don't design or work with databases today.
What is one important concept that developers and other software professionals ought to know about databases?
The very first thing developers should know about databases is this: what are databases for? Not how do they work, nor how do you build one, nor even how do you write code to retrieve or update the data in a database. But what are they for?
Unfortunately, the answer to this one is a moving target. In the heydey of databases, the 1970s through the early 1990s, databases were for the sharing of data. If you were using a database, and you weren't sharing data you were either involved in an academic project or you were wasting resources, including yourself. Setting up a database and taming a DBMS were such monumental tasks that the payback, in terms of data exploited multiple times, had to be huge to match the investment.
Over the last 15 years, databases have come to be used for storing the persistent data associated with just one application. Building a database for MySQL, or Access, or SQL Server has become so routine that databases have become almost a routine part of an ordinary application. Sometimes, that initial limited mission gets pushed upward by mission creep, as the real value of the data becomes apparent. Unfortunately, databases that were designed with a single purpose in mind often fail dramatically when they begin to be pushed into a role that's enterprise wide and mission critical.
The second thing developers need to learn about databases is the whole data centric view of the world. The data centric world view is more different from the process centric world view than anything most developers have ever learned. Compared to this gap, the gap between structured programming and object oriented programming is relatively small.
The third thing developers need to learn, at least in an overview, is data modeling, including conceptual data modeling, logical data modeling, and physical data modeling.
Conceptual data modeling is really requirements analysis from a data centric point of view.
Logical data modeling is generally the application of a specific data model to the requirements discovered in conceptual data modeling. The relational model is used far more than any other specific model, and developers need to learn the relational model for sure. Designing a powerful and relevant relational model for a nontrivial requirement is not a trivial task. You can't build good SQL tables if you misunderstand the relational model.
Physical data modeling is generally DBMS specific, and doesn't need to be learned in much detail, unless the developer is also the database builder or the DBA. What developers do need to understand is the extent to which physical database design can be separated from logical database design, and the extent to which producing a high speed database can be accomplished just by tweaking the physical design.
The next thing developers need to learn is that while speed (performance) is important, other measures of design goodness are even more important, such as the ability to revise and extend the scope of the database down the road, or simplicity of programming.
Finally, anybody who messes with databases needs to understand that the value of data often outlasts the system that captured it.
Whew!
Good question. The following are some thoughts in no particular order:
Normalization, to at least the second normal form, is essential.
Referential integrity is also essential, with proper cascading delete and update considerations.
Good and proper use of check constraints. Let the database do as much work as possible.
Don't scatter business logic in both the database and middle tier code. Pick one or the other, preferably in middle tier code.
Decide on a consistent approach for primary keys and clustered keys.
Don't over index. Choose your indexes wisely.
Consistent table and column naming. Pick a standard and stick to it.
Limit the number of columns in the database that will accept null values.
Don't get carried away with triggers. They have their use but can complicate things in a hurry.
Be careful with UDFs. They are great but can cause performance problems when you're not aware how often they might get called in a query.
Get Celko's book on database design. The man is arrogant but knows his stuff.
First, developers need to understand that there is something to know about databases. They're not just magic devices where you put in the SQL and get out result sets, but rather very complicated pieces of software with their own logic and quirks.
Second, that there are different database setups for different purposes. You do not want a developer making historical reports off an on-line transactional database if there's a data warehouse available.
Third, developers need to understand basic SQL, including joins.
Past this, it depends on how closely the developers are involved. I've worked in jobs where I was developer and de facto DBA, where the DBAs were just down the aisle, and where the DBAs are off in their own area. (I dislike the third.) Assuming the developers are involved in database design:
They need to understand basic normalization, at least the first three normal forms. Anything beyond that, get a DBA. For those with any experience with US courtrooms (and random television shows count here), there's the mnemonic "Depend on the key, the whole key, and nothing but the key, so help you Codd."
They need to have a clue about indexes, by which I mean they should have some idea what indexes they need and how they're likely to affect performance. This means not having useless indices, but not being afraid to add them to assist queries. Anything further (like the balance) should be left for the DBA.
They need to understand the need for data integrity, and be able to point to where they're verifying the data and what they're doing if they find problems. This doesn't have to be in the database (where it will be difficult to issue a meaningful error message for the user), but has to be somewhere.
They should have the basic knowledge of how to get a plan, and how to read it in general (at least enough to tell whether the algorithms are efficient or not).
They should know vaguely what a trigger is, what a view is, and that it's possible to partition pieces of databases. They don't need any sort of details, but they need to know to ask the DBA about these things.
They should of course know not to meddle with production data, or production code, or anything like that, and they should know that all source code goes into a VCS.
I've doubtless forgotten something, but the average developer need not be a DBA, provided there is a real DBA at hand.
Basic Indexing
I'm always shocked to see a table or an entire database with no indexes, or arbitrary/useless indexes. Even if you're not designing the database and just have to write some queries, it's still vital to understand, at a minimum:
What's indexed in your database and what's not:
The difference between types of scans, how they're chosen, and how the way you write a query can influence that choice;
The concept of coverage (why you shouldn't just write SELECT *);
The difference between a clustered and non-clustered index;
Why more/bigger indexes are not necessarily better;
Why you should try to avoid wrapping filter columns in functions.
Designers should also be aware of common index anti-patterns, for example:
The Access anti-pattern (indexing every column, one by one)
The Catch-All anti-pattern (one massive index on all or most columns, apparently created under the mistaken impression that it would speed up every conceivable query involving any of those columns).
The quality of a database's indexing - and whether or not you take advantage of it with the queries you write - accounts for by far the most significant chunk of performance. 9 out of 10 questions posted on SO and other forums complaining about poor performance invariably turn out to be due to poor indexing or a non-sargable expression.
Normalization
It always depresses me to see somebody struggling to write an excessively complicated query that would have been completely straightforward with a normalized design ("Show me total sales per region.").
If you understand this at the outset and design accordingly, you'll save yourself a lot of pain later. It's easy to denormalize for performance after you've normalized; it's not so easy to normalize a database that wasn't designed that way from the start.
At the very least, you should know what 3NF is and how to get there. With most transactional databases, this is a very good balance between making queries easy to write and maintaining good performance.
How Indexes Work
It's probably not the most important, but for sure the most underestimated topic.
The problem with indexing is that SQL tutorials usually don't mention them at all and that all the toy examples work without any index.
Even more experienced developers can write fairly good (and complex) SQL without knowing more about indexes than "An index makes the query fast".
That's because SQL databases do a very good job working as black-box:
Tell me what you need (gimme SQL), I'll take care of it.
And that works perfectly to retrieve the correct results. The author of the SQL doesn't need to know what the system is doing behind the scenes--until everything becomes sooo slooooow.....
That's when indexing becomes a topic. But that's usually very late and somebody (some company?) is already suffering from a real problem.
That's why I believe indexing is the No. 1 topic not to forget when working with databases. Unfortunately, it is very easy to forget it.
Disclaimer
The arguments are borrowed from the preface of my free eBook "Use The Index, Luke". I am spending quite a lot of my time explaining how indexes work and how to use them properly.
I just want to point out an observation - that is that it seems that the majority of responses assume database is interchangeable with relational databases. There are also object databases, flat file databases. It is important to asses the needs of the of the software project at hand. From a programmer perspective the database decision can be delayed until later. Data modeling on the other hand can be achieved early on and lead to much success.
I think data modeling is a key component and is a relatively old concept yet it is one that has been forgotten by many in the software industry. Data modeling, especially conceptual modeling, can reveal the functional behavior of a system and can be relied on as a road map for development.
On the other hand, the type of database required can be determined based on many different factors to include environment, user volume, and available local hardware such as harddrive space.
Avoiding SQL injection and how to secure your database
Every developer should know that this is false: "Profiling a database operation is completely different from profiling code."
There is a clear Big-O in the traditional sense. When you do an EXPLAIN PLAN (or the equivalent) you're seeing the algorithm. Some algorithms involve nested loops and are O( n ^ 2 ). Other algorithms involve B-tree lookups and are O( n log n ).
This is very, very serious. It's central to understanding why indexes matter. It's central to understanding the speed-normalization-denormalization tradeoffs. It's central to understanding why a data warehouse uses a star-schema which is not normalized for transactional updates.
If you're unclear on the algorithm being used do the following. Stop. Explain the Query Execution plan. Adjust indexes accordingly.
Also, the corollary: More Indexes are Not Better.
Sometimes an index focused on one operation will slow other operations down. Depending on the ratio of the two operations, adding an index may have good effects, no overall impact, or be detrimental to overall performance.
I think every developer should understand that databases require a different paradigm.
When writing a query to get at your data, a set-based approach is needed. Many people with an interative background struggle with this. And yet, when they embrace it, they can achieve far better results, even though the solution may not be the one that first presented itself in their iterative-focussed minds.
Excellent question. Let's see, first no one should consider querying a datbase who does not thoroughly understand joins. That's like driving a car without knowing where the steering wheel and brakes are. You also need to know datatypes and how to choose the best one.
Another thing that developers should understand is that there are three things you should have in mind when designing a database:
Data integrity - if the data can't be relied on you essentially have no data - this means do not put required logic in the application as many other sources may touch the database. Constraints, foreign keys and sometimes triggers are necessary to data integrity. Don't fail to use them because you don't like them or don't want to be bothered to understand them.
Performance - it is very hard to refactor a poorly performing database and performance should be considered from the start. There are many ways to do the same query and some are known to be faster almost always, it is short-sighted not to learn and use these ways. Read some books on performance tuning before designing queries or database structures.
Security - this data is the life-blood of your company, it also frequently contains personal information that can be stolen. Learn to protect your data from SQL injection attacks and fraud and identity theft.
When querying a database, it is easy to get the wrong answer. Make sure you understand your data model thoroughly. Remember often actual decisions are made based on the data your query returns. When it is wrong, the wrong business decisions are made. You can kill a company from bad queries or loose a big customer. Data has meaning, developers often seem to forget that.
Data almost never goes away, think in terms of storing data over time instead of just how to get it in today. That database that worked fine when it had a hundred thousand records, may not be so nice in ten years. Applications rarely last as long as data. This is one reason why designing for performance is critical.
Your database will probaly need fields that the application doesn't need to see. Things like GUIDs for replication, date inserted fields. etc. You also may need to store history of changes and who made them when and be able to restore bad changes from this storehouse. Think about how you intend to do this before you come ask a web site how to fix the problem where you forgot to put a where clause on an update and updated the whole table.
Never develop in a newer version of a database than the production version. Never, never, never develop directly against a production database.
If you don't have a database administrator, make sure someone is making backups and knows how to restore them and has tested restoring them.
Database code is code, there is no excuse for not keeping it in source control just like the rest of your code.
Evolutionary Database Design. http://martinfowler.com/articles/evodb.html
These agile methodologies make database change process manageable, predictable and testable.
Developers should know, what it takes to refactor a production database in terms of version control, continious integration and automated testing.
Evolutionary Database Design process has administrative aspects, for example a column is to be dropped after some life time period in all databases of this codebase.
At least know, that Database Refactoring concept and methodologies exist.
http://www.agiledata.org/essays/databaseRefactoringCatalog.html
Classification and process description makes it possible to implement tooling for these refactorings too.
About the following comment to Walter M.'s answer:
"Very well written! And the historical perspective is great for people who weren't doing database work at that time (i.e. me)".
The historical perspective is in a certain sense absolutely crucial. "Those who forget history, are doomed to repeat it.". Cfr XML repeating the hierarchical mistakes of the past, graph databases repeating the network mistakes of the past, OO systems forcing the hierarchical model upon users while everybody with even just a tenth of a brain should know that the hierarchical model is not suitable for general-purpose representation of the real world, etcetera, etcetera.
As for the question itself:
Every database developer should know that "Relational" is not equal to "SQL". Then they would understand why they are being let down so abysmally by the DBMS vendors, and why they should be telling those same vendors to come up with better stuff (e.g. DBMS's that are truly relational) if they want to go on sucking hilarious amounts of money out of their customers for such crappy software).
And every database developer should know everything about the relational algebra. Then there would no longer be a single developer left who had to post these stupid "I don't know how to do my job and want someone else to do it for me" questions on Stack Overflow anymore.
From my experience with relational databases, every developer should know:
- The different data types:
Using the correct type for the correct job will make your DB design more robust, your queries faster and your life easier.
- Learn about 1xM and MxM:
This is the bread and butter for relational databases. You need to understand one-to-many and many-to-many relations and apply then when appropriate.
- "K.I.S.S." principle applies to the DB as well:
Simplicity always works best. Provided you have studied how DB work, you will avoid unnecessary complexity which will lead to maintenance and speed problems.
- Indices:
It's not enough if you know what they are. You need to understand when to used them and when not to.
also:
Boolean algebra is your friend
Images: Don't store them on the DB. Don't ask why.
Test DELETE with SELECT
I would like everyone, both DBAs and developer/designer/architects, to better understand how to properly model a business domain, and how to map/translate that business domain model into both a normalized database logical model, an optimized physical model, and an appropriate object oriented class model, each one of which is (can be) different, for various reasons, and understand when, why, and how they are (or should be) different from one another.
I would say strong basic SQL skills. I've seen a lot of developers so far who know a little about databases but are always asking for tips about how to formulate a quite simple query. Queries are not always that easy and simple. You do have to use multiple joins (inner, left, etc.) when querying a well normalized database.
I think a lot of the technical details have been covered here and I don't want to add to them. The one thing I want to say is more social than technical, don't fall for the "DBA knowing the best" trap as an application developer.
If you are having performance issues with query take ownership of the problem too. Do your own research and push for the DBAs to explain what's happening and how their solutions are addressing the problem.
Come up with your own suggestions too after you have done the research. That is, I try to find a cooperative solution to the problem rather than leaving database issues to the DBAs.
Simple respect.
It's not just a repository
You probably don't know better than the vendor or the DBAs
You won't support it at 3 a.m. with senior managers shouting at you
Consider Denormalization as a possible angel, not the devil, and also consider NoSQL databases as an alternative to relational databases.
Also, I think the Entity-Relation model is a must-know for every developper even if you don't design databases. It'll let you understand thoroughly what's your database all about.
Never insert data with the wrong text encoding.
Once your database becomes polluted with multiple encodings, the best you can do is apply some kind combination of heuristics and manual labor.
Aside from syntax and conceptual options they employ (such as joins, triggers, and stored procedures), one thing that will be critical for every developer employing a database is this:
Know how your engine is going to perform the query you are writing with specificity.
The reason I think this is so important is simply production stability. You should know how your code performs so you're not stopping all execution in your thread while you wait for a long function to complete, so why would you not want to know how your query will affect the database, your program, and perhaps even the server?
This is actually something that has hit my R&D team more times than missing semicolons or the like. The presumtion is the query will execute quickly because it does on their development system with only a few thousand rows in the tables. Even if the production database is the same size, it is more than likely going to be used a lot more, and thus suffer from other constraints like multiple users accessing it at the same time, or something going wrong with another query elsewhere, thus delaying the result of this query.
Even simple things like how joins affect performance of a query are invaluable in production. There are many features of many database engines that make things easier conceptually, but may introduce gotchas in performance if not thought of clearly.
Know your database engine execution process and plan for it.
For a middle-of-the-road professional developer who uses databases a lot (writing/maintaining queries daily or almost daily), I think the expectation should be the same as any other field: You wrote one in college.
Every C++ geek wrote a string class in college. Every graphics geek wrote a raytracer in college. Every web geek wrote interactive websites (usually before we had "web frameworks") in college. Every hardware nerd (and even software nerds) built a CPU in college. Every physician dissected an entire cadaver in college, even if she's only going to take my blood pressure and tell me my cholesterol is too high today. Why would databases be any different?
Unfortunately, they do seem different, today, for some reason. People want .NET programmers to know how strings work in C, but the internals of your RDBMS shouldn't concern you too much.
It's virtually impossible to get the same level of understanding from just reading about them, or even working your way down from the top. But if you start at the bottom and understand each piece, then it's relatively easy to figure out the specifics for your database. Even things that lots of database geeks can't seem to grok, like when to use a non-relational database.
Maybe that's a bit strict, especially if you didn't study computer science in college. I'll tone it down some: You could write one today, completely, from scratch. I don't care if you know the specifics of how the PostgreSQL query optimizer works, but if you know enough to write one yourself, it probably won't be too different from what they did. And you know, it's really not that hard to write a basic one.
The order of columns in a non-unique index is important.
The first column should be the column that has the most variability in its content (i.e. cardinality).
This is to aid SQL Server ability to create useful statistics in how to use the index at runtime.
Understand the tools that you use to program the database!!!
I wasted so much time trying to understand why my code was mysteriously failing.
If you're using .NET, for example, you need to know how to properly use the objects in the System.Data.SqlClient namespace. You need to know how to manage your SqlConnection objects to make sure they are opened, closed, and when necessary, disposed properly.
You need to know that when you use a SqlDataReader, it is necessary to close it separately from your SqlConnection. You need to understand how to keep connections open when appropriate to how to minimize the number of hits to the database (because they are relatively expensive in terms of computing time).
Basic SQL skills.
Indexing.
Deal with different incarnations of DATE/ TIME/ TIMESTAMP.
JDBC driver documentation for the platform you are using.
Deal with binary data types (CLOB, BLOB, etc.)
For some projects, and Object-Oriented model is better.
For other projects, a Relational model is better.
The impedance mismatch problem, and know the common deficiencies or ORMs.
RDBMS Compatibility
Look if it is needed to run the application in more than one RDBMS. If yes, it might be necessary to:
avoid RDBMS SQL extensions
eliminate triggers and store procedures
follow strict SQL standards
convert field data types
change transaction isolation levels
Otherwise, these questions should be treated separately and different versions (or configurations) of the application would be developed.
Don't depend on the order of rows returned by an SQL query.
Three (things) is the magic number:
Your database needs version control too.
Cursors are slow and you probably don't need them.
Triggers are evil*
*almost always
I am in the early stages of developing a database-driven system and the largest part of the system revolves around an inheritance type of relationship. There is a parent entity with about 10 columns and there will be about 10 child entities inheriting from the parent. Each child entity will have about 10 columns. I thought it made sense to give the parent entity its own table and give each of the children their own tables - a table-per-subclass structure.
Today, my users requested to see the structure of the system I created. They balked at the idea of the table-per-subclass structure. They would prefer one big ~100 column table because it would be easier for them to perform their own custom queries.
Should I consider denormalizing the database for the sake of the users?
Absolutely not. You can always create a view later to show them what they want to see.
They are effectively asking for a report.
You could give them access to a view containing all the fields they require... that way you don't mess up your data model.
No. Structure the data properly and if the users need the a denormalized view of the data create it as a VIEW in the database.
Alternatively, consider that perhaps an RDBMS is not the appropriate storage tool for this project.
They are the users and not the programmers of the system for a reason. Provide a separate interface for their queries. Power users like this can both be helpful and a pain to deal with. Just explain you need the database designed a certain way so you can do your job, period. Once that is accomplished you and provide other means to make querying easier.
What do they know!? You could argue that users shouldn't even be having direct access to a database in the first place.
Doing that leaves you open to massive performance issues, just because a couple of users are running ridiculous queries.
How about if you created a VIEW in the format your users wanted while still maintaining a properly normalized table?
Aside from a lot of the technical reasons for or against your users' proposition, you need to be on same page in communicating the consequences of various scenarious and (more importantly) the costs of those consequences. If the users are your clients and they are paying you to do a job, explain that their awful "proposed" ideas may cost them more money in development time, additional hardware resources, etc.
Hopefully you can explain it in such a way that shows your expertise and why your idea is a much better value to your users in the long run.
As everyone more or less mentioned, that way lies madness, and you can always build a view.
If you just can't get them to come around on this point, consider showing them this thread and the number of pros who weighed in saying that the users are meddling with things that they don't fully understand, and the impact will be an undermined foundation.
A big part of the developer's craft is the feel for what won't work out long term, and the rules of normalization are almost canonical in that respect. There are situations where you need to denormalize (data warehouses, etc) but this doesn't sound like one of them!
It also sounds as though you may have a particularly troubling brand of user on your hand -- the amatuer developer who thinks they could do your job better themselves if only they had the time. This may or may not help, but I've found that those types respond well to presentation -- a few times now I've found that if I dress sharp and show a little bit of force in my personality, it helps them feel like I'm an expert and prevents a bunch of problems before they start.
I would strongly recommend coming up with an answer that doesn't involve someone running direct reports against your database. The moment that happens, your DB structure is set in stone and you can basically consider it legacy.
A view is a good start, but later on you'll probably want to structure this as an export, to decouple further. Of course, then you'll encounter someone who wants "real time" data. Proper business analysis usually reveals this to be unnecessary. Actual real time requirements are not best handled through reporting systems.
Just to be clear: I'd personally favour the table per subclass approach, but I don't think it's actually as big an issue as the direct reporting off transaction tables is going to be.
I would opt for a view (as others have suggested) or an inline table-valued function (the benefits of this is you require parameters - like an date range or a customer account - which can help to stop users from querying without any limits on the problem space) first. An inline TVF is really a parametrized view and is far closer to a view in terms of how the engine treats them than it is to a multi-statement table valued function or a scalar function, which can perform incredibly poorly.
However, in some cases, this can impact production performance if the view is complex or intensive. With poorly written ad hoc user queries, it can also cause locks to persist longer or be escalated further than they would on a better built query. It is also possible for users to misinterpret an E-R data model and produce multiplied numbers in cases where there are many-to-one or many-to-many relationships. The next option might be to materialize these views with indexes or make tables and keep them updated, which gets us closer to my next option...
So, given those drawbacks of the view option and already thinking of mitigating it by starting to make copies of data, the next option I would consider is to have a separate read-only (for these users) version of the data which is structured differently. Typically, I would first look at a Kimball-style star schema. You do not need to have a full-fledged time-consistent data warehouse. Of course, that's an option, but you could simply keep a reporting model up to date with data. Star-schemas are a special form of denormalization and are particularly good for numerical reporting, and a given star should not be able to be abused by users accidentally. You can keep the star up to date in a number of ways, including triggers, scheduled jobs, etc. They can be very fast for reporting needs and run on the same production installation - perhaps on a separate instance if not just a separate database.
Although such a solution may require you to effectively more than double your storage requirements, when compared with other practices it might be a really good option if you understand your data well and don't mind having two models - one for transactions and one for analysis (note that you will already start to have this logical separation anyway with the use of a the simplest first option of view).
Some architects will often double their servers and use the SAME model with some kind of replication in order to provide a reporting server which is indexed more heavily or differently. Such a second server doesn't impact production transactions with reporting requirements and can be kept up to date fairly easily. There will only be one model, but of course, this has the same usability problems with allowing users ad hoc access to the underlying model only, without the performance affects, since they get their own playground.
There are a lot of ways to skin these cats. Good luck.
The customer is always right. However, the customer is likely to back down when you convert their requirement into dollars and cents. A 100 column table will require extra dev time to write the code that does what the database would do automatically with the proper implementation. Further, their support costs will be higher since more code means more problems and lower ease of debugging.
I'm going to play devil's advocate here and say that both solutions sound like poor approximations of the actual data. There's a reason that object-oriented programming languages don't tend to be implemented with either of these data models, and it's not because Codd's 1970 ideas about relations were the ideal system for storing and querying object-oriented data structures. :-)
Remember that SQL was originally designed as a user interface language (that's why it looks vaguely like English and not at all like other languages of that era: Algol, C, APL, Prolog). The only reasons I've heard for not exposing a SQL database to users today are security (they could take down the server!) and usability (who wants to write SQL when you can clicky clicky?), but if it's their server and they want to, then why not let them?
Given that "the largest part of the system revolves around an inheritance type of relationship", then I'd seriously consider a database that lets me represent that natively, either Postgres (if SQL is important) or a native object database (which are awesome to work with, if you don't need SQL compatibility).
Finally, remember that every engineering decision is a tradeoff. By "sticking to your guns" (as somebody else proposed), you're implicitly saying the value of your users' desires are zero. Don't ask SO for a correct answer to this, because we don't know what your users want to do with your data (or even what your data is, or who your users are). Go tell them why you want a many-tables solution, and then work out a solution with them that's acceptable to both of you.
You've implemented Class Table Inheritance and they're asking for Single Table Inheritance. Both designs are valid in certain situations.
You might want to get a copy of Martin Fowler's Patterns of Enterprise Application Architecture to read more about the advantages and disadvantages of each design. That book is a classic reference to have on your bookshelf, in any case.
I am supposed to perform ETL where source is a large and badly designed sql 2k database and a a better designed sql 2k5 database. I think SSIS is the way to go. Can anyone suggest a to-do list or a checklist or things to watchout for so that I dont forget anything? How should I approach this so that it does not bite me in the rear later on.
Some general ETL tips
Consider organising it by
destination (for example, all the
code to produce the Customer
dimension lives in the same module, regardless of source).
This is sometimes known as
Subject-oriented ETL. It makes
finding stuff much easier and will
increase the maintainability of your
code.
If the SQL2000 database is a mess,
you will probably find that SSIS
data flows are a clumsy way to deal
with the data. As a rule, ETL tools
scale poorly with complexity;
something like half of all data
warehouse projects in finance
companies are done with stored
procedure code as an explicit
architectural decision - for precisely this reason. If you have
to put a large amount of code in
sprocs, consider putting all of the
code in sprocs.
For a system involving lots of complex scrubbing or transformations, a 100% sproc approach is far more maintainable as it is the only feasible way to put all of the transformations and business logic in one place. With mixed ETL/sproc systems, you have to look in multiple places to track, troubleshoot, debug or change the whole transformation.
The sweet spot of ETL tools is on systems where you have a larger number of data sources with relatively simple transformations.
Make the code testable, so you can
pick apart the components and test
in isolation. Code that can only be executed from within the middle of a complex data flow in an ETL tool is much harder to test.
Make the data extract dumb with no
business logic, and copy into a
staging area. If you have business
logic spread across the extract and
transform layers, you will have
transformations that cannot be tested
in isolation and make it hard to
track down bugs. If the transform is
running from a staging area you
reduce the hard dependency on the
source system, again enhancing
testability. This is a particular win on sproc-based architectures as it allows an almost completely homogeneous code base.
Build a generic slowly-changing
dimension handler or use one off the
shelf if available. This makes it
easier to unit test this
functionality. If this can be unit
tested, the system testing does not
have to test all of the corner cases,
merely whether the data presented to
it is correct. This is not as complex as it sounds - The last one I wrote was about 600 or 700 lines of T-SQL code. The same goes for any generic scrubbing functions.
Load incrementally if possible.
Instrument your code - have it make log entries, possibly recording diagnostics such as check totals or counts. Without this, troubleshooting is next to impossible. Also, assertion checking is a good way to think of error handling for this (does row count in a equal row count in b, is A:B relationship really 1:1).
Use synthetic keys. Using natural keys from the source systems ties your system to the data sources, and makes it difficult to add extra sources. The keys and relationships in the system should always line up - no nulls. For errors, 'not recorded', make a specific 'error' or 'not recorded' entries in the dimension table and match to them.
If you build an Operational Data Store (the subject of many a religious war) do not recycle the ODS keys in the star schemas. By all means join on ODS keys to construct dimensions, but match on a natural key. This allows you to arbitrarily drop and recreate the ODS - possibly changing its structure - without disturbing the star schemas. Having this capability is a real maintenance win, as you can change ODS structure or do a brute-force re-deployment of the ODS at any point.
Points 1-2 and 4-5 mean that you can build a system where all of the code for any given subsystem (e.g. a single dimension or fact table) lives in one and only one place in the system. This type of architecture is also better for larger numbers of data sources.
Point 3 is a counterpoint to point 2. Basically the choice between SQL and ETL tooling is a function of transformation complexity and number of source systems. The simpler the data and larger the number of data sources, the stronger the case for a tools-based approach. The more complex the data, the stronger the case for moving to an architecture based on stored procedures. Generally it's better to exclusively or almost exclusively use one or the other but not both.
Point 6 makes your system easier to test. Testing SCD's or any change based functionality is fiddly, as you have to be able to present more than one version of the source data to the system. If you move the change management functionality into infrastructure code, you can test it in isolation with test data sets. This is a win in testing, as it reduces the complexity of your system testing requirements.
Point 7 is a general performance tip that you will need to observe for large data volumes. Note that you may only need incremental loading for some parts of a system; for smaller reference tables and dimensions you may not need it.
Point 8 is germane to any headless process. If it goes tits up during the night, you want some fighting chance of seeing what went wrong the next day. If the code doesn't properly log what's going on and catch errors, you will have a much harder job troubleshooting it.
Point 9 gives the data warehouse a life of its own. You can easily add and drop source systems when the warehouse has its own keys. Warehouse keys are also necessary to implement slowly changing dimensions.
Point 10 is a maintenance and deployment win, as the ODS can be re-structured if you need to add new systems or change the cardinality of a record. It also means that a dimension can be loaded from more than one place in the ODS (think: adding manual accounting adjustments) without a dependency on the ODS keys.
I have experience with ETL processes pulling data from 200+ distributed databases to a central database on a daily, weekly, monthly and yearly basis. It is a massive amount of data and there are many issues we have had specific to our situation. But as I see it, there are several items to think about regardless of the situation:
Make sure that you take file locks into consideration, both on the source and destination side. Making sure that other processes do not have the files locked (and removing those locks if necessary and it makes sense) is important.
locking the files for yourself. Make sure, especially on the source that you lock the files while pulling out the data so that you do not get halfway updated data.
if at all possible, pull deltas, not all of the data. Get a copy of the data and then pull only rows that have changed instead of everything. The larger your data set the more important this becomes. Look at journals and triggers if you have to, but as it becomes more important to have this data on a certain basis, this is probably the number one advice I would give you. Even if it adds a significant amount of time to the project.
execution log. make sure you know when it worked and when it didn't, and throwing specific errors in the process can really help in debugging.
document, document, document. If you build this right, you are going to build it and then not think about it for a long time. But you can be guaranteed, you or someone else will need to come back to it at some point to enhance it or do a bug fix. Documentation is key in these situations.
HTH, ill update this if I think of anything else.
Well i'm developing an ETL for the company where i am.
We are working with SSIS.
Using the api to generate and build our own dtsx packages.
SSIS it's not friendly for managing errors. Sometimes you get an "OleDb Error" that could have a lot of different meanings depeding on the context.
Read the API Documentation (they don't say much).
Some links to help you out starting there:
http://technet.microsoft.com/de-de/library/ms135932(SQL.90).aspx
http://msdn.microsoft.com/en-us/library/ms345167.aspx
http://msdn.microsoft.com/en-us/library/ms403356.aspx
http://www.codeproject.com/KB/database/SSISProgramming.aspx?display=PrintAll&fid=382208&df=90&mpp=25&noise=3&sort=Position&view=Quick&fr=26&select=2551674
http://www.codeproject.com/KB/database/foreachadossis.aspx
http://wiki.sqlis.com/default.aspx/SQLISWiki/ComponentErrorCodes.html
http://www.new.facebook.com/inbox/readmessage.php?t=1041904880323#/home.php?ref=logo
http://technet.microsoft.com/en-us/library/ms187670.aspx
http://msdn.microsoft.com/ja-jp/library/microsoft.sqlserver.dts.runtime.foreachloop.foreachenumerator.aspx
http://www.sqlis.com/post/Handling-different-row-types-in-the-same-file.aspx
http://technet.microsoft.com/en-us/library/ms135967(SQL.90).aspx
http://msdn.microsoft.com/en-us/library/ms137709(SQL.90).aspx
http://msdn.microsoft.com/en-us/library/ms345164(SQL.90).aspx
http://msdn.microsoft.com/en-us/library/ms141232.aspx
http://www.microsoft.com/technet/prodtechnol/sql/2005/ssisperf.mspx
http://www.ivolva.com/ssis_code_generator.html
http://www.ivolva.com/ssis_wizards.html
http://www.codeplex.com/MSFTISProdSamples
http://www.experts-exchange.com/Microsoft/Development/MS-SQL-Server/SSIS/Q_23972361.html
http://forums.microsoft.com/MSDN/MigratedForum.aspx?siteid=1&PostID=1404157
http://msdn.microsoft.com/en-us/library/aa719592(VS.71).aspx
http://forums.microsoft.com/MSDN/MigratedForum.aspx?siteid=1&ForumID=80
http://blogs.conchango.com/jamiethomson/archive/2005/06/11/SSIS_3A00_-Custom-Logging-Using-Event-Handlers.aspx
http://blogs.conchango.com/jamiethomson/archive/2007/03/13/SSIS_3A00_-Property-Paths-syntax.aspx
http://search.live.com/results.aspx?q=%s&go=Buscar&form=QBJK&q1=macro%3Ajamiet.ssis
http://toddmcdermid.blogspot.com/2008/09/using-performupgrade.html?showComment=1224715020000
http://msdn.microsoft.com/en-us/library/ms136082.aspx
http://support.microsoft.com/kb/839279/en-us
Sorry for the "spam", but they are very useful to me.
We're doing a huge ETL (moving a client from legacy AS400 apps to Oracle EBS), and we actually have a process that (with modifications) I can recommend:
Identify the critical target
tables/fields.
Identify the critical
source tables/fields.
Work with the
business users to map source to
target.
Analyze the source data for
quality issues.
Determine who's
responsible for data quality issues
identified.
Have responsible parties
clean up the data in the source.
Develop the actual ETL based on the
information from steps 1 - 3.
The trickiest steps are 2 & 3 in my experience - it's sometimes difficult to get the business users to correctly identify all the bits they need in one pass, and can be even harder to properly identify exactly where the data is coming from (though that may have something to do with cryptic file and field names that I'm seeing!). However, this process should help you avoid major misses.
This thread is old, but I want to draw your attention to ConcernedOfTunbridgeWells' answer. It is incredibly good advice, on all points. I could reiterate a few, but that would diminish the rest, and they all deserve close study.