For starters, why use SQL Server CE? I think it’s an easy way of implementing a high supported and functional database solution for various types of applications although this article will focus on using SQL Server CE in a small Windows forms application. To begin with you’ll need to download SQL Server CE from the Microsoft website.

In the past I think I’ve gone a bit overboard with the length of my posts so I’m going to attempt to keep this one short. The application I used this code in is pretty well split up and uses a number of different layers but you don’t need to go this far to use SQL CE with your applications. The code below is a summary only and covers the key things I think you need to know – feel free to ask any questions you need to.

I’m also aware that the code below *will* need modification before you can just throw it into your own applications but, as mentioned above, please post questions on this post if there is anything you need clarification on.

Creating the database ::

1
2
3
4
5

string connStr = String.Format("Data Source = C:\temp\database.sdf");
using (SqlCeEngine engine = new SqlCeEngine(connStr))
{
   engine.CreateDatabase();
}

Running a SQL command that doesn’t return a result set (note that this is, in some cases, where you would execute SQL to create the database structure after using the code above to create the database file itself) ::

1
2
3
4
5
6
7
8
9
10

using (SqlCeConnection conn = new SqlCeConnection(connStr))
{
   using (SqlCeCommand cmd = new SqlCeCommand())
   {
      cmd.Connection = conn;
      conn.Open();
 
      cmd.CommandText = "SELECT * FROM table_name";
   }
}

Getting a record from the database ::

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

using (var conn = new SqlCeConnection(connStr))
{
   conn.Open();
 
   string query = "SELECT * FROM Applications WHERE id = @Id";
 
   using (var cmd = new SqlCeCommand(query, conn))
   {
      // appId is a variable passed to the method that uses this code sample
      cmd.Parameters.AddWithValue("@Id", appId);
      var reader = cmd.ExecuteResultSet(ResultSetOptions.Scrollable | ResultSetOptions.Updatable);
      if (reader.HasRows)
      {
         while (reader.Read())
         {
            int Id = Convert.ToInt32(reader["Id"]);
            // more code here to deal with other columns in the record that was selected from the database, if any
         }
      }
   }
}

That’s about as basic as it gets but I’m also in the process of writing another article about a basic application that uses this code in a “real-world” situation.

Unattended installations of SQL Server can be as involved as you want – pretty much every option can be specified in the configuration file. For this example I’m going to perform a relatively basic installation with minimal customisation so you can see the process at work. The installation will specify most of what I think are the critical options you’ll be interested in. Note that for this example I’m using the x86 (32-bit) version with advanced services.

Assumptions:

1. You’ve downloaded the x86 (32-bit) version of SQL Server 2008 Express from the Microsoft Microsoft SQL Server 2008 Express website
2. Your system meets the requirements for installing SQL Server 2008
3. You are installing from C:\Install\SQL2008 (you can change this as necessary but please make sure you update the paths in the examples below)
4. You’ll save your configuration file as C:\Install\sql-2008-express.ini

Once you’ve satisfied those requirements follow the steps below.

1. Run the executable you downloaded with the /x parameter (e.g. C:\Install\SQL2008\SQLEXPRADV_x86_ENU.exe
2. Choose a location to extract the installation files to.
3. Create a new text file that will become the unattended installation configuration file. Leave it empty for now.

Now that you’ve got a nicely extracted copy of the installation files you can carry on. The above steps were necessary for what’s coming up.

So, you can now create your own configuration file using Microsoft’s documentation or you can use the version I’ve included in this article.

Be careful with the Microsoft documentation as there is an error on the page called How to: Install SQL Server 2008 from the Command Prompt. It mentions a parameter called /BROWSERSVRACCOUNT but this should be /BROWSERSVCACCOUNT. If you get this wrong the installation WILL fail.

You can download the configuration file I’ve made below. The changes I’ve made are as follows (you may need to change these to match your settings).

• Set the QUIETSIMPLE parameter to “True”.
• Set the MEDIASOURCE parameter to the appropriate installation path.
• Added a parameter called SAPWD and set it to the strong ‘sa’ password.
• Set the INDICATEPROGRESS parameter to “True”.
• Set the SQLSYSADMINACCOUNTS value to “MYPC\Administrator”.

Once you have your answer file all nicely setup and ready to use you’ll need to tell the setup program how to use it.

From a command prompt (or Start > Run if you’re into that) run the following command. Remember to change the file and path names to the ones that match your system.

If you’ve done everything right this will start the installation process and you’ll end up with a SQL Server 2008 instance called SQLExpress running on your system. Easy!

Click here to download a ZIP file that contains the full configuration file I’ve used in this article plus a script that does an unattended installation based on the configuration file.

The instance in question has the default name of SQLEXPRESS. Obviously this means that to connect to the server you need to use SERVER\SQLEXPRESS – that works fine from Management Studio and from the Java application installed on this particular server. You can read about the problems I had with this installation in my article called Apache Tomcat 6 – .WAR won’t deploy.

When using SQLCMD.EXE from the command line the full error message looks like this:

Some Googling suggested that this problem could be caused by one of 2 things. Firstly, that the SQL Browser service isn’t running – on my server it was running. Secondly, named pipes aren’t enabled in the SQL Server Configuration Manager … on mine they are. Hmmm.

I’ve had problems in the past with named instances of SQL Server so I had a look at the properties of the named pipes configuration and, sure enough, it said \\.\pipe\MSSQL$SQLEXPRESS\sql\query. That *looks* ok, right? Yes but it’s the cause of the problem.

I changed the named pipe to the following.

Voila! SQLCMD.EXE now works from the command line meaning my scheduled backups now work too. Not being a full-time DBA I’d put this down to a bit of a fluke on my part although I then found an article by Jesse Johnston that confirmed the same steps worked in their situation too.

For a start, the INSERT statement’s syntax was perfectly fine – I tested it using a new query from within Microsoft Access. The fields in question were called ‘email’ and ‘identifier’ and it turned ‘email’ is a reserved word in Microsoft Access. One of the original SQL statements looked something like this:

That is one of the statements that caused the error. Because of the Microsoft Access reserved words you must enclose the column names in square brackets, like this:

Enclosing the column names in square brackets is apparently best practice anyway (as found in many MS SQL Server queries) but nonetheless … problem solved!

The question posted was the following:

The answer is as follows (with the exception of the hide advanced configuration options – it’s not necessary to worry about that during an unattended install).

Hope that helps someone.

One of the features of the full and very expensive version of SQL Server 2005 is the ability to run scheduled backups using SQL Maintenance Plans. SQL Server requires SSIS (SQL Server Integration Services) for these to be available – unfortunately SQL Server Express doesn’t include SSIS. There is a way to get scheduled backups for SQL Server Express though.

Before you begin you’ll need to install SQL Server Management Studio Express.

Part 1 – Create a backup script

- Open SQL Server Management Studio Express and login as a user with sufficient permission to access the database(s) you want to backup. You can do this as ‘sa’ if you want.
- Expand the name of your server or instance, expand Databases and select the database you want to backup.
- Right-click the database name, select Tasks and then select Back Up.
- Set the options you require for the backup. For example you might want to change the location where the backup files get created, edit backup set retention period etc.
- Once you’ve set the options necessary click the Script drop-down option at the top of the backup window and select either ‘Script Action to File’, or, if you want to see/edit the contents of the script before saving it, select ‘Script Action to New Query Window’. I find it’s best to script the backup to a new query window because you can test it before saving.
- Save the created script when you’re finished editing or enter an appropriate filename immediately if you selected ‘Script Action to File’.
- If you saved the script, open it up – we’re going to test it before making it ‘live’.
- Once you have the script open click the checkmark/tick button just to make absolutely sure the script is valid (it’s SQL Server-generated so it’d better be!).
- Click the ‘Execute’ button if you want to run a test. If you are backing up a very large database this can take a while and you won’t get a whole lot of feedback while the backup is being run. When the backup finishes the results will be displayed in the Results pane, successful or not. If it fails, figure out why (e.g. check for sufficient disk space, permissions for the folder you’re backing up to etc).

The complete script for one of my databases is shown at the end of this article.

Part 2 – Schedule the backup

This part is simply about creating a Windows scheduled task – pretty easy really.

If you installed SQL Server Express into the default location the syntax for the command you need to run is as follows (note the Binn directory used, not Bin!)

C:\Program Files\Microsoft SQL Server\90\Tools\Binn\SQLCMD.EXE -s <database> -i <sql script>

For example:

C:\Program Files\Microsoft SQL Server\90\Tools\Binn\SQLCMD.EXE -s my_database -i c:\scripts\backup_my_database.sql

So, let’s setup the scheduled task.

- Open Control Panel and select Scheduled Tasks
- Click Add Scheduled Task, and browse to C:\Program Files\Microsoft SQL Server\90\Tools\Binn\SQLCMD.EXE when prompted for the command to run.
- Select an appropriate frequency for the job and click Next.
- Enter when you want the job to run and click Next.
- Enter the credentials for the job. These credentials should have permissions to both the database and the directory the database will be backed up to. CLick Next.
- Check the box that says open advanced properties for this task when I click Finish then click Finish.
- Change the ‘Run’ command to be the complete command above including the name of the database and the script you want to run, e.g. C:\Program Files\Microsoft SQL Server\90\Tools\Binn\SQLCMD.EXE -s my_database -i c:\scripts\backup_my_database.sql.
- Click OK and confirm the password for the user running the scheduled task if you are prompted to do so.

That’s it, you’re done. If you want to make absolutely sure it’s going to run as expected you can right-click the scheduled task and select ‘Run’ – this will do make sure everything is configured ok. Make sure you see a SQL backup file in the selected location after this. If so, everything should be ok for the scheduled run.

Don’t forget to add the backup directory to your offsite or tape backups (if you are doing them, AND YOU SHOULD BE!)

Here is a complete backup script for one of my databases. It’s configured to backup everything in a database called ‘prod’ and retain the backups for a period of 7 days after which they will expire. This helps limit the amount of disk space used by the backups. This script is also set to verify the backup when done.

BACKUP DATABASE [prod] TO  DISK = N'C:\Backups\SQL\prod.bak' WITH  RETAINDAYS = 7, NOFORMAT, NOINIT,  NAME = N'prod-Full Database Backup', SKIP, NOREWIND, NOUNLOAD,  STATS = 10
GO
declare @backupSetId as int
select @backupSetId = position from msdb..backupset where database_name=N'prod' and backup_set_id=(select max(backup_set_id) from msdb..backupset where database_name=N'prod' )
if @backupSetId is null begin raiserror(N'Verify failed. Backup information for database ''prod'' not found.', 16, 1) end
RESTORE VERIFYONLY FROM  DISK = N'C:\Backups\SQL\prod.bak' WITH  FILE = @backupSetId,  NOUNLOAD,  NOREWIND
GO

This is due to the new Vista feature called User Account Control or UAC. The fix is pretty simple and there are 2 different ways of implementing it.

The second method is to disable UAC completely but I wouldn’t recommend it.

First, some background info. I studied The National Diploma in Business Computing ages ago but started work before I completed it. A big mistake on my part as it turns out that pretty much every job you go for these days wants to see some evidence of the ability to complete studies, projects … anything that demonstrates commitment really.

Before I continue I’ll just say that this article is a bit of a brain dump. It’s as much for me as it is for anyone else as I find it easier to remember stuff if I write something about it. If you get something out of this article, cool. Please keep in mind that until I started this paper I could spell Oracle but that was about it … apologies to any Oracle gurus out there who think my ramblings are a waste of time.

Recently I looked into how much of my Diploma I had to go before it was finished. Up until then I thought I had maybe 5 or 6 papers to do but I found that I only had 2! The first paper, 157.221 (Information Systems Analysis), is a pre-requisite for 157.223 (Information Systems Design). 157.221 was already underway so I was a bit gutted as I thought I wouldn’t be able to complete the necessary papers in time (157.223 is being phased out at the end of 2007).

So anyway, I did some more digging and the Information Systems staff at Massey University of Wellington agreed to grant me the right to study 157.223 without having completed 157.221. This was after an interview with two of the Information Systems staff members to assess my ability to complete 157.223 without studying the material in 157.221. Sweet!

So now that the background is out of the way, what’s this Oracle thing then?

Ok, I’m not really going to harp on about what Oracle is other than to say it’s currently one of best known database systems around. The 157.223 course concentrates on using Oracle to teach the concepts of database design and implementation (Oracle-style of course).

So far we’ve only had 1 lab session on Oracle but it’s pretty cool … HUGE, but cool. The Oracle people have been kind enough to give anyone the right to download and install Oracle 10g, the previous version of Oracle Database (the latest is Oracle 11g). The only requirement is that you don’t use the downloaded software for commercial purposes.

In the first lab session we hacked about with Oracle Designer, the tool that supplies utilities that allow you to start from no more than an idea and work your way towards a fully-functional database system. The recommended steps of database design are all covered by the Oracle Designer utilities. In other words you can get the GUI tool to create, amongst other things, your ER diagrams and table structures, both obviously critical parts of the final design. It seems to me that even if you’ve had very little experience with database design you could probably grab Oracle 10g and throw together a fully-functioning system pretty quickly … yes you’ll need to read the manuals but that shouldn’t be a problem for anyone wanting to do it properly.

I’m typing this during what is supposed to be Lab #2 but unfortunately that isn’t happening due to a “glitch in The Matrix” … in other words the IT Services staff here have mucked something up and Oracle isn’t installed on our lab machines anymore. Damn. Interestingly they use Altiris Deployment Server, a product I’ve had reasonably extensive experience with. In the time they spent finding out that things were broken they probably could’ve got Oracle installed on all the lab machines … maybe they should’ve asked me. Just kidding.

Anyway, where was I. Oh yeah, Oracle. We’ve also spent some time using the Database Design Transformer, a tool that can be used to create your database’s initial, first-cut design from the analysis data model – a representation of the business’s information needs. The result of this process is to, for example, describe basic transformation rules and predict the resulting transformations.

After we had our first-cut design sorted out we then had to use the Design Editor to refine that design even further. The Design Editor is described in our text as “the primary Oracle Designer tool for design and build tasks.” So far I’d agree … hehe.

In the lab scheduled for today we were supposed to do a whole load more with refinement of our design but as mentioned earlier there’s no Oracle around for me to use right now so I’ll have to skip writing about that … doh. Maybe next week Oracle will be reinstalled and I can actually write something interesting.

Bye!

- SQL Authentication [ login to the server using a SQL Server login ]
- Windows Authentication [ login to the server using a Windows/Active Directory login ]
- Mixed Mode Authentication [ login to the server using either of the two types above ]

There are a couple of ways to accomplish this. Most people, sometimes to their demise, configure their servers for “Mixed Mode” authentication, i.e. allow both SQL and Windows logins access to the server. Sometimes, however, this is a requirement, for example if the application connecting to MSDE simple can’t use Windows authentication. To enable mixed mode authentication you can specify the option during setup, as follows :

Note the use of the SAPWD parameter to specify a strong ‘sa’ password at install time.

If you’ve already installed MSDE and don’t want to reinstall it you can modify the authentication mode by using Enterprise Manager and changing the properties of the SQL instance you’re working with or you can use the registry editor (the registry edit method may be the only option available to you if you don’t have any administration tools like Enterprise Manager installed). If we assume you’re modifying a default instance of SQL Server the registry key you need to change is :

The modes are as follows :

0 = SQL Authentication
1 = Windows Authentication
2 = Mixed Mode Authentication

After you’ve made this change you will need to stop and restart the MSDE services.

USE REGISTRY EDITOR AT YOUR OWN RISK!

Here is the Microsoft article re MSDE Security & Authentication.

Click the following link to read my post about fixing “Access Denied” errors when connecting to MSDE remotely after installation.

where is the password you want for the ‘sa’ account. For example :

Click here for more information on the parameters you can use when running MSDE setup.

So anyway, once you’ve done that you might find you can’t connect to your nice new shiny MSDE server over the network due to an error that looks like the following :

This is because by default there are no remote protocols enabled, i.e. you can’t connect remotely. To fix this you can either hack up the registry but we’re not going to do that here. Instead we’re going to use the SVRNETCN.EXE utility that comes with SQL Server. Unless you’ve installed MSDE as a named instance this executable is available in the following path :

Run this and you’ll get something that looks like the following.

For typical networks/installations using TCP/IP will be sufficient. If this is ok for you select TCP/IP from the list of ‘Disabled Protocols’ on the left side of the screen and click the ‘Enable >>’ button to enable the protocol. Click OK and accept the message that comes up saying you need to restart SQL Server before the change will take effect.

From Control Panel | Administrative Tools | Services you’ll then need to find the service for the instance of SQL that you’re working with (by default this is called “MSSQLSERVER”) and restart it … right-click the service name and select “Restart”. If SQLSERVERAGENT is running you’ll be asked if you want to restart this too – you’ll need to say Yes to continue.

After this you should be able to use Query Analyser or Enterprise Manager to administer your MSDE installation using the ‘sa’ account (note that MSDE doesn’t come with these tools as standard though).

If you want to login to your MSDE server using Windows Authentication you have to enable that too … I’ll post about that shortly (there are registry changes involved if you don’t run setup with the appropriate switches at install time).