Related
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
We don’t allow questions seeking recommendations for books, tools, software libraries, and more. You can edit the question so it can be answered with facts and citations.
Closed 6 years ago.
Improve this question
For a current project I'm creating a data model. Are there any sources where I can find "best practices" for a good data model? Good means flexible, efficient, with good performance, style, ... Some example questions would be "naming of columns", "what data should be normalized", or "which attributes should be exported into an own table". The source should be a book :-)
Personally I think you should read a book on performance tuning before beginning to model a database. The right design can make a world of difference. If you are not expert in performance tuning, you aren't qualified to design a database.
These books are Database specific, here is one for SQl Server.
http://www.amazon.com/Server-Performance-Tuning-Distilled-Experts/dp/1430219025/ref=sr_1_1?s=books&ie=UTF8&qid=1313603282&sr=1-1
Another book that you should read before starting to design is about antipatterns. Always good to know what you should avoid doing.
http://www.amazon.com/SQL-Antipatterns-Programming-Pragmatic-Programmers/dp/1934356557/ref=sr_1_1?s=books&ie=UTF8&qid=1313603622&sr=1-1
Do not get stuck in the trap of designing for flexibility. People use that as a way to get out of doing the work to design correctly and flexible databases almost always perform badly. If more than 5% of your database design depends on flexibility, you haven't modeled correctly in my opinion. All the worst COTS products I've had to work with were designed for flexibility first.
Any decent database book will discuss normalization. You can also find that information easily on the web. Be sure to actually create FK/PK relationships.
As far as naming columns, pick a standard and stick with it consistently. Consistency is more important than the actual standard. Don't name columns ID (see SQL antipatterns book). Use the same name and datatypes if columns are going to be in several different tables. What you are going for is to not have to use functions to do joins because of datatype mismatches.
Always remember that databases can (and will) be changed outside the application. Anything that is needed for data integrity must be in the database not the application code. The data will be there long after the application has been replaced.
The most important things for database design:
Thorough definition of the data needed (including correct datatypes)
and the relationships between pieces of data (including correct normalization)
data integrity
performance
security
consistency (of datatypes, naming standards etc.)
The best book I've read on the design of database systems was "An Introduction to Database Systems". Joe Celko's SQL for Smarties books are also worth reading.
Assuming you're building an application and not just a database, and assuming you're using an Object Oriented language, Applying UML and Patterns by Craig Larman has a good discussion on mapping databases to objects.
In terms of defining "good", in my experience "maintainable" is probably top of the list. Maintainability in database design means many things, such as sticking to conventions - I often recommend http://justinsomnia.org/2003/04/essential-database-naming-conventions-and-style/. Normalization is another obvious maintainability strategy. I often recommend being generous with column types - it's hard to change an application if you find out that postal codes in different countries are longer than in the US. I often recommend using views to abstract complex data relations away for less experienced developers.
A key thing with maintainability is the ability to test and deploy. It's worth reading up about Continuous Database Integration (http://www.codeproject.com/KB/architecture/Database_CI.aspx) - whilst not strictly associated with the design of the database schema, it's important context.
As for performance - I believe you should design for maintainability first, and only design for performance if you know you have a problem. Sometimes, you know in advance that performance will be a major problem - designing a database for Facebook (or Stack Exchange), designing a database with huge amounts of data (terabytes and up), or huge numbers of users. Most systems don't fall into that camp - so I recommend regular performance tests, with representative data, to find if you have a problem, and only tune when you can prove you have to. Many performance optimizations are at the expense of maintainability - denormalization, for instance.
Oh, and in general, avoid triggers and stored procedures if you can. That's just my opinion, though...
Even though it is not a book I recommend to read Query evaluation techniques for large databases. It gives a background on query processing which largely influences your schema design, especially for data intensive (e.g., analytical) workloads. It is less hands-on but I believe every database designer should read it at least once :-).
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.
My employer has a database committee and we've been discussing different platforms. Thus far we've had people present on SqlLite, MySQL, and PostreSql. We currently use the Microsoft stack so we're all quite familiar with Microsoft Sql Server.
As a part of this comparison I thought it would be interesting to create a small reference application for each database platform to explore the details involved in working with it.
First: does that idea make sense or is the comparison require going beyond the scope of a trivial sample application?
Second: I would imagine each reference application having a discrete but small set of requirements that fulfill many of the scenarios we run into on a regular basis. Here is what I have so far, what else can be added to the list but still keep the application small enough to be built in a very limited timespan?
Connectivity from the application layer
Tools for database administration
Process of creating a schema (small "s" schema, tables/views/functions other objects)
Simple CRUD (Create, Retrieve, Update, Delete)
Transaction support
Third: has anyone gone through this process, what are your findings?
Does that idea make sense or is the
comparison require going beyond the
scope of a trivial sample application?
I don't think it's a good idea. Most of the things that will really affect you are long term database management issues, and how the database management system you choose can handle those things.
You could be tempted in the short term with things like "I found out in 3 seconds how to do this with XYZ database management system". Now, I'm not saying support is not important; quite the contrary. But finding an answer in google in 3 seconds means that you got an answer to a simple question. How quickly, if ever, can you find an answer to a challenging problem?
A short list (not exhaustive) of important things to consider are:
backup and recovery -- at both logical level and physical level
good support for functions (or stored procedures), triggers, various SQL query constructs
APIs that allow real extensibility -- these things can get you out of tough situations and allow you to solve problems in creative ways. You'd be surprised what can be accomplished with user-defined types and functions. How do the user-defined types interact with the indexing system?
SQL standard support -- doesn't trump everything else, but if support is lacking in a few areas, really consider why it is lacking, what the workarounds are, and what are the costs of those workarounds.
A powerful executor that offers a range of fundamental algorithms (e.g. hash join, merge join, etc.) and indexing structures (btree, hash, maybe a full text option, etc.). If it's missing some algorithms or index structures, consider the types of questions that the database will be inefficient at answering. Note: I don't just mean "slow" here; the wrong algorithm can easily be worse by orders of magnitude.
Can the type system reasonably represent your business? If the set of types available is incredibly weak, you will have a mess. Representing everything as strings is kind of like assembly programming (untyped), and you will have a mess.
A trivial application won't show you any of those things. Simple things are simple to solve. If you have a "database committee" then your company cares about its data, and you should take the responsibility seriously. You need to make sure that you can develop applications on it easily with the results you and your developers expect; and when you run into problems you need to have access to a powerful system and quality support that can get you through it.
Actually learning capabilities of each RDMS is more crucial. Because it depends on the application. If you need spatial data capabilities PostGIS with PostgreSQL is better than MySQL. If you need easy replication, high availability features MySQL seems better. Also there are license issues. A link for comparison here. All has strengths and weaknesses. First get the requirements of your project or projects than compare it with list the features of the RDMSs you pick and decide which one to go.
I don't think you need to test the simple CRUD stuff, it's hard to imagine a vendor that doesn't support the basics.
Firstly, you're going beyond the scope of a sample app, in my humble opinion.
Secondly, I'd pick the one most appropriate to the tool or application you wish to develop. For example, are schemas and transactions relevant for a database that stores a single-user app configuration?
Thirdly, I've worked with Access, SQL Server, SQLite, MySQL, PostgreSQL and Oracle, and they all have their place. If you're in the MS space, go with SQL Server (and don't forget Express). There are also ADO.NET ways to talk to the others in my list. It depends on what you want.
Frankly, I doubt an arbitrarily-defined simple application would be likely to really highlight the differences between database engines. I think you'd be better to read the advertising literature for the various engines to see what they claim as their strong points. Then consider which of these matter to you, and construct tests specifically designed to verify claims that you care about.
For example, here are pros and cons of database engines I've used the most that have mattered to me. I don't claim this is an exhaustive list, but it may give you an idea of things to think about:
MySQL: Note: MySQL has two main engines internally: MyISAM and InnoDB. I have never used the InnoDB.
Pros: Fast. Free to cheap depending on how you're using it. Very convenient and easy-to-use commands for managing the schema. Some very useful extensions to the SQL standard, like "insert ... on duplicate".
Cons: The MyISAM engine does not support transactions, i.e. there's no rollback. MyISAM engine does not manage foreign keys for you. (InnoDB does not have these drawbacks, but as I say, I've never used it, so I can't comment much further.) Many deviations from SQL standards.
Oracle: Pros: Fast. Generally good conformance to SQL standards. My brother works for Oracle so if you buy there you'll be helping support my family. (Okay, maybe that's not an important pro for you ...)
Cons: Difficult to install and manage. Expensive.
Postgres: Pros: Very high conformance to SQL standards. Free. Very good "explain" plans.
Cons: Relatively slow. Optimizer is easily confused on complex queries. Some awkwardness in modifying existing tables.
Access: Pros: Easy to install and manage. Very easy to use schema management. Built-in data entry tools and query builder for quick-and-dirty stuff. Cheap.
Cons: Slow. Unreliable with multiple users.
I think that you can investigate Firebird too
This is an extract of Firebird-General on yahoogroups and I find it quite objective
Our natural audience is developers who
want to package and sell proprietary
applications. Firebird is easier to
package and install than Postgres;
more capable than SQLite; and doesn't
charge a royalty like MySQL.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 4 years ago.
Improve this question
If someone knew little about databases and wanted to learn about them from scratch, which database would you recommend learning with and why?
MySQL seems ubiquitous, but are there others that are more modern that have learned lessons from the past, or others that are simply nicer or more logical to work with?
Universal compatibility/libraries is not a big concern, unless it is something truly obscure. Mac (Unix) compatibility is a must.
If you just want to be learning the SQL language, and not database administration, I would recommend working with SQLite. If you're on a Mac, it should already be installed. It is a much simpler system than most RDBMSes; there is no server to set up, and client to connect to the server. There are no directories of cryptic files, or anything of the sort. To get started, you can just type:
sqlite3 mydatabase.db
And start working with it. It's so much lighter weight and easier to set up and use than the other database systems that I think it's a good choice for a beginner.
Now, SQLite is a fairly small and lightweight language. If you need be getting into any kind of really complex queries and data mining, I would recommend PostgreSQL. It has a fairly advanced query optimizer, and a pretty long list of SQL features.
And if you want to learn a database as something to use for back-end storage for web programming or something of the sort, MySQL is what I'd choose. It's ubiquitous, supported by almost any web host, and it's pretty fast for very simple queries and updates, which is generally what you need for a web system. It has some real gotchas to avoid when setting it up; you have to choose between several different storage engines, and it can take a lot of work to convince it to actually work with Unicode data. But it's good to learn mostly for its ubiquity.
From what I've seen (at least on the web), MySQL and PostgreSQL are the most ubiquitous free database systems. If you're considering learning one of them, check this comparison out.
You may also want to consider learning SQLite, a "self-contained, embeddable, zero-configuration SQL database engine." It's really easy to get up and going, stored in a single file, and as its description says, has no complicated configuration. SQLite has proved enormously popular as a persistent data store for local apps on the desktop/iPhone. If you're going down this route on a Mac/iPhone, you may also want to check out Core Data, which is an abstraction layer Apple developed on top of SQLite(but can work with pretty much any DB), to simplify working with a database. As a bonus, Core Data includes a nice GUI for forming relationships and entities. You can check out this tutorial for more information.
If you really, truly, want to "learn from scratch", then theory is the first thing to learn. And that means : NOT products, not any. Not DB2, not MySQL, not oracle, not any-of-them.
Hugh Darwen has a freely available e-book entitled "An Introduction to Relational Database Theory". The material is quite "accessible" and quite unlike most other theory textbooks. It's also the accompanying textbook for his university course on database technology.
Chris Date has several books, of which "Introduction to database systems" is the most comprehensive, also the standard textbook in the field, but maybe a little too abstract for some.
If you think that all you need is "just to know a product" and that you can do equally well "without all that theory", then in that case, please disregard this response, because the wording of your question is dishonest.
Sad thing about databases is that each and every one works bit differently. I would most likely pick MySQL first and play with it a bit. Then get PostgreSQL and do the same.
If you need to use databases in corporate environment then I would aim to test also Oracle and SQL Server which both have express versions that can be installed free for yourself.
http://www.microsoft.com/express/sql/download/
http://www.oracle.com/technology/software/products/database/index.html
At start all databases are more or less confusing but I would pick MySQL as first because it can do most of the basic functionalities and has a lot of help available.
I'd go with mysql. It's easy to setup, easy to mess around with via with mysql client, and it's well documented. If you're just starting out, you probably won't need most of the features offered by other databases, like stored procedures and the like.
First of all, MySQL is both ubiquitous and modern.
ANSI SQL is more or less the same in all RDBMs, so you can learn any of them and you'll be good.
Once you've mastered ANSI SQL, then all you've got left is the localized solutions for each one of them, which won't be portable to other systems - and so, totally discouraged to use, unless they simplify your tasks in a way justifying it.
MySQL, PostgreSQL, SQLite - pick one. PostgreSQL is more like Oracle, and in my opinion a bit more mature. It's had stored procedures, triggers, and referential integrity longer than MySQL has. I'll admit that I have both installed, but I use MySQL more often because it's quick and easy.
But do be aware that non-SQL alternatives are out there and growing in importance. BigTable, object databases like db4o, are worth being aware of. "No SQL" is out there.
If you are just getting your feet wet, MYSQL is a great one to start with. Easy install on any platform, great community support and lots of free tools to work with (SQLYog is a favorite of mine).
I agree that theory is very important. Depending on how you learn best, digging in and tinkering may be the thing to do before you try to absorb 40 years of thought on relational systems.
Codd and Date are legends in the field and can help you understand the broader points of relational theory, but are hard to absorb before you have context for the topic.
If you are looking for more pragmatic/immediate guidance, I'd suggest a book like "Databases for Mere Mortals" and anything written by Joe Celko.
Once you get comfortable with the basics, there are lots of other platforms to explore as well. As mentioned above, SQLLite and PostGres are two other great choices for the Mac OS.
If you want to learn SQL : the best way is to choose database who implement more features of the SQL standard. So I would recommand Firebird or PostgreSQL
I might be "sidetracking" a bit with this answer, but I think we're in the same situation!
Check out the "The Manga Guide to Databases"! I haven't read it myself yet, but it's on the way in the mail as we speak! I've heard good things about it from friends and colleges, and it's got some good reviews as well. Albeit a bit "controversial," it's supposed to be a fun and surprisingly in-depth introduction to fundamental techniques and principles!
Alex wrote: "Reading a textbook without incrementally testing your knowledge on an actual database is not going to produce good results for the majority of people."
My book and my university course both use Dave Vooorhis's Rel for that very purpose.
Hugh
A database in the cloud: Amazon EC2, Google App Engine or Microsoft Azure
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 9 years ago.
Improve this question
How do SQL developers go about keeping up on current techniques and trends in the SQL world? Are there any blogs, books, articles, techniques, etc that are being used to keep up to date and in the know?
There are a lot of opportunities out their for OO, procedural, and functional programmers to take part in a variety of open source projects, but it seems to me that the FOSS avenue is a bit more closed for SQL developers.
Thoughts?
Find challenging questions that test your TRANSACT-SQL knowledge ... personally I enjoy Joe Celko's SQL Puzzles and Answers.
Joe Celko's SQL Puzzles and Answers http://ecx.images-amazon.com/images/I/51DTJ099P7L._SL500_BO2,204,203,200_PIsitb-dp-500-arrow,TopRight,45,-64_OU01_AA240_SH20_.jpg
The thing about the SQL language is that it is pretty much a static target. Pretty soon you are looking at increasing your understanding of set theory and the problem domain itself rather than the details of the language.
The real meat is on either side of the language, in either the databases themselves (how to store, retrieve, and organize large data sets) or in the applications (with ORMs and such)
I skimmed the answers and apparently nobody has mentioned Stephane Faroult's work.
I strongly suggest you should consider "The Art of SQL" (http://www.amazon.com/Art-SQL-Stephane-Faroult/dp/0596008945), I found it really interesting and - amazingly enough, even fun to read.
How about http://sqlblog.com/
I improve by analyzing slow and complex queries and looking for ways to improve them. This can be done in SQL Server by analyzing the Query Plan tools and looking for bottlenecks. Also I find the Visual Quickstart Guide guide to be good for quick reference.
Joe Celko's SQL Puzzles and Answers and SQL for Smarties are the two best generic SQL books out there. Both are great sources to give you ideas for that tricky problem you used to think you needed a cursor or some client library to accomplish. For any truly interested SQL geek, these book are also pretty good for casual reading rather than as a mere desk reference. Two thumbs up.
While not a SQL Server expert, in general I find that community based events are great ways to keep up on current patterns. The underlying result of participating in a community of developers/DBAs/Marketing Pros/insert profession here is that you are learning new thought patterns and excercising critical thought. This is a great way to grow as whatever professional you are.
There aren't current techniques and trends in SQL. There's only that stuff you should already know but don't. The proper way to learn that stuff, is pain... so much pain.
Join a mailing list for the DB flavour you use...or lurk on stackoverflow ;)
Most "current" stuff is not SQL itself, but how the database stores the information, and how to retrieve it more quickly. Check out this other thread: What are some references, lessons and or best practices for SQL optimization training
The only real bleeding edge is in query planning, index structures, sort algorithms, things like that, not the SQL itself.
The fact that you asked this question is already a good sign. Avoiding complacency is "piece of advice #1". There is no substitute for writing and optimizing SQL. Practical use is the best way to stay sharp, but there is a risk of a "forest for the trees" scenario, where we tend to use what is comfortable and familiar. Trying new tactics, examining new approaches, and looking for new ways to train our brains to think about sets, SQL, relational theory, and staying on top of new developments in the particular dialects we employ are all hallmarks of good SQL developers.
There are many good blogs out there these days. I work mostly in the Microsoft arena, so I like SQLTeam.com.
Usenet is a good place to hang out and make a contribution. There are many SQL-related newsgroups. Often, you will find that working on someone else's problem helps you learn a new tactic or forces you to research a dusty corner of the language that you do not encounter every day. ISPs seem destined to shut all of the Usenet down, though, because of nefarious use, so this one may be going the way of the Dodo bird.
Also, some IRC servers have a vibrant sql channels where you can make the same sort of a difference (just take a thick skin with you).
Lastly, this very website might be another place to hang, where you can read over the answers to difficult questions, see how that might apply in your own world, practice the techniques, and internalize them. Contribute too, because seeing how others vote your solutions up or down is 100% pure honest feedback.
Of course, there are many wonderful books out there, too. Anything by Celko is a winner, and on the SQL Server side, Kalen Delaney and Ron Soukup have written some winners.
Best thing I've run into is working on other people's SQL code. Especially legacy business code. You want to test your skills against something, start changing some "voodoo code" that no one else understands. :)
Beyond that, I just try to keep an eye on changes with new releases of SQL and see if there's anything I can take advantage of.
gleam tips when using phpmyadmin
it's nice and verbose
Here is one with some interesting SSIS information.
http://blogs.conchango.com/jamiethomson/default.aspx
There is also some good information in the Wiki here:
http://wiki.lessthandot.com/index.php/Main_Page
For those who say SQL never changes, SQL Server 2005 and 2008 have some huge changes in the T-SQl that will help solve some difficult problems that were horrible to do in SQL Server 2000 and are much easier once you learn the new syntax, so yes there is stuff to keep up with.
Also performance tuning and SSIS are extremely complex subjects with much to learn.
I do find that developers who choose not to learn advanced SQL skills tend to write poorly performing SQL code and once the number of records grows in their databases, the applications they wrote tend to become glacially slow and very difficult to fix at that point. Right now I'm working with developers to fix some bad code they wrote that is causing timeouts on the site on virtually every query. Obviously, this is now an emergency and it would have been easy to write the code in a more efficeint manner at the start if the developer had better SQL skills.
I've never heard of the term "SQL developer." SQL should be a skill in your toolbox, like sorting, whatever framework you like, JavaScript, and so on. The best way to continue to improve your SQL skills to continue using it.
As a developer, and not a DBA, I keep an eye on various developer resources, and that often is DB related, but I don't specifically 'try to keep up'.
I know plenty, but I also know that there is so much more that there is to know. And in every project I have to learn something new. And every project also involves me taking a different approach to a similar task I'd encountered in the past.
Should I ever get to the stage where I think I'm doing the same things all the time, perhaps I'll make a concious effort to take specific steps. But currently, and for the foreseeable future, I'm learning organically, on-the-job, and as my projects dictate.
Reading:
Books - Celko (also read across to some Oracle-biased books)
Blogs - the above mentioned, plus SSWUG
Webinars and Conference - Best way to keep up with vendor-specific stuff like
SSIS/SSRS/SSAS
Practice:
Improving code (mine and others)
Refactoring
Mentoring/training other developers
Honestly, it's one of those things that you just get better at with time. Read as much as you can to know what's possible. Some things will take a while to really understand. I was scared off by sub queries for a long time until I pretty much had no choice but to use them.
When you get more experiance and need to do more complex things, you will just learn your way.
SqlServerCentral - great source of articles, scripts, advice
Unfortunately to access the articles you need to register (it is free though)
I guess one thing they could learn from StackOverflow is to remove login barrier
We've written a full tutorial, and you can test your SQL skills at a separate site (also created by our me, in the interests of full declaration).
SQL developers, or DBAs?
Aside from learning different dialects of SQL (Oracle, SQL Server, etc) in your day to day work, SQL doesn't actually change all that much. Sure you can bring in more advanced concepts as you develop your skills, work out where to implement stored procedures, etc, but in the end it's just SQL. The most important thing is to get your schema correct and maintainable.
Now administering the databases is a whole different thing, with a range of tools, and the database software itself getting updated every few years. Oracle at least have newsletters and websites and magazines that presumably include lots of information and examples and best practice scenarios.
To be honest, I don't see much a need for extreme SQL skills. Once I can create transactions (for DB consistency) and basic triggers (for cross-table consistency), I'm usually fine keeping program logic... in the program, and not putting it into whatever database I'm using. I've not found much depth to SQL worth investigating for a lifetime, unlike general programming, which keeps expanding in depth.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 5 years ago.
Improve this question
Modern database systems today come with loads of features. And you would agree with me that to learn one database you must unlearn the concepts you learned in another database. For example, each database would implement locking differently than others. So to carry the concepts of one database to another would be a recipe for failure. And there could be other examples where two databases would perform very very differently.
So while developing the database driven systems should the programmers need to know the database in detail so that they code for performance? I don't think it would be appropriate to have the DBA called for performance later as his job is to only maintain the database and help out the developer in case of an emergency but not on a regular basis.
What do you think should be the extent the developer needs to gain an insight into the database?
I think these are the most important things (from most important to least, IMO):
SQL (obviously) - It helps to know how to at least do basic queries, aggregates (sum(), etc), and inner joins
Normalization - DB design skills are an major requirement
Locking Model/MVCC - Its nice to have at least a basic grasp of how your databases manage row locking (or use MVCC to accomplish similar goals with optimistic locking)
ACID compliance, Txns - Please know how these work and interact
Indexing - While I don't think that you need to be an expert in tablespaces, placing data on separate drives for optimal performance, and other minutiae, it does help to have a high level knowledge of how index scans work vs. tablescans. It also helps to be able to read a query plan and understand why it might be choosing one over the other.
Basic Tools - You'll probably find yourself wanting to copy production data to a test environment at some point, so knowing the basics of how to restore/backup your database will be important.
Fortunately, there are some great FOSS and free commercial databases out there today that can be used to learn quite a bit about db fundamentals.
I think a developer should have a fairly good grasp of how their database system works, not matter which one it is. When making design and architecture decisions, they need to understand the possible implications when it comes to the database.
Personally, I think you should know how databases work as well as the relational model and the rhetoric behind it, including all forms of normalization (even though I rarely see a need to go beyond third normal form). The core concepts of the relational model do not change from relational database to relational database - implementation may, but so what?
Developers that don't understand the rationale behind database normalization, indexes, etc. are going to suffer if they ever work on a non-trivial project.
I think it really depends on your job. If you are a developer in a large company with dedicated DBAs then maybe you don't need to know much, but if you are in a small company then it may be really helpful knowing more about databases. In small companies you may wear more than one hat.
It cannot hurt to know more in any situation.
It certainly can't hurt to be familiar with relational database theory, and have a good working knowledge of the standard SQL syntax, as well as knowing what stored procedures, triggers, views, and indexes are. Obviously it's not terribly important to learn the database-specific extensions to SQL (T-SQL, PL/SQL, etc) until you start working with that database.
I think it's important to have a basic understand of databses when developing database applications just like it's important to have an understanding of the hardware your your software runs on. You don't have to be an expert, but you shouldn't be totally ignorant of anything your software interacts with.
That said, you probably shouldn't need to do much SQL as an application developer. Most of the interaction with the database should be done through stored procedures developed by the DBA, I'm not a big fan of including SQL code in your application code. If your queries are in stored procedures, then the DBA can change the implementation of the stored procedure, or even the database schema, and so long as the result is the same it doesn't require any changes to your application code.
If you are uncertain about how to best access the database you should be using tried and tested solutions like the application blocks from Microsoft - http://msdn.microsoft.com/en-us/library/cc309504.aspx. They can also prove helpful to you by examining how that code is implemented.
Basic things about Sql queries are must. then you can develop simple system. but when you are going to implement Complex systems you should know Normalization, Procedures, Functions, etc.