|
|
YOUR FEEDBACK
SOA World Conference
Virtualization Conference $200 Savings Expire May 16, 2008... – Register Today! Did you read today's front page stories & breaking news?
SYS-CON.TV SYS-CON.TV WEBCASTS |
TOP COLDFUSION LINKS Journeyman CF
Reconfiguring Remote Databases via SQL
By: Charlie Arehart
Digg This!
Here's a dilemma faced by nearly every developer of Access databases who's got the database stored on a remote CF server. How do you reconfigure the database if you want to add, change or remove columns, tables or indexes? If it's remote, you can't very well use the Access interface to do so. So you've probably bitten the bullet and resorted to simply downloading the database, changing it locally, then uploading it back to the server. It's a brute force method, and it does indeed work, but at a price - especially compared to a much faster, safer and generally more productive approach. This month we look at how to perform such database configurations remotely through SQL. You'll still use CFQUERY (though there are some issues to be aware of), but you'll use different SQL commands that may be new to you.
What's Wrong with the Brute Force Approach?
Of course, this brute force approach is limited to file system databases in which the changed database can be downloaded/uploaded, which typically means we're referring to Access. There are plenty of arguments against the long-term viability of running a production application in Access, but it's what many are forced or choose to use. Users of databases like SQL Server, Sybase and Oracle have probably been using SQL for a long time, since the brute force approach simply wasn't available to them. And those enterprise-class DBMSs also typically offer tools to administer a remote database with a properly configured network connection to the remote server. But what if you don't have such a tool, or can't get that "proper network connection" arranged? Then this SQL approach may be important, and new to you, too. Finally, note that SQL Server 7 databases are now based on files (rather than "devices") that can be downloaded/uploaded in a similar fashion, so novice SQL Server developers may be applying the brute force approach as well.
About the SQL Approach First decide what you want to do (add/change/remove columns/indexes/tables, etc.), then determine the SQL DDL (Data Definition Language) statement to use and determine the appropriate parameters for the statements (such as data types available when adding columns, which may vary depending on the database management system). We'll discuss the available actions and corresponding SQL DDL statements later. There are even tools to help build this DDL. (The SQL statements you're more familiar with, such as SELECT, INSERT and UPDATE, are referred to as DML - Data Manipulation Language. They're both still SQL, and are executed the same way in CF. They're just classified separately.) You'd be wise to make your next step a test of the set of DDL statements on a local copy of the database - which needn't necessarily have the latest data from the remote version. You just want to make sure you haven't made a mistake in your SQL. If it works locally, it'll almost certainly work remotely - though not necessarily. If the database drivers on your local and remote system aren't the same, the code could respond differently. Finally, before attempting to execute the tested set of SQL remotely, it would probably be wise to create a backup of the database on the remote server. While that may seem to negate the time savings of the brute force approach and open up some of the integrity issues referred to before, keep in mind that making a copy on the remote server itself (via operating system file management commands or CFFILE) will be very fast, or at least orders of magnitude faster than downloading/uploading the file even over a fast Internet connection. I lay these steps out now, before focusing on the all-important SQL DDL statements, because I want you to appreciate all the steps in the approach before getting overly excited about it. The SQL is pretty easy to use, but it's also easy to make a mistake. Approaching this process in a cavalier manner would be dangerous to your data, your users and possibly your job! Whenever you're dealing with reconfiguring your databases, you must exercise supreme caution. It's not really that what we're discussing here is any more dangerous than making the edits via the brute force approach. You can make a mistake in either case. But since you're issuing commands rather than using a graphical user interface, there's a little more potential for making a mistake. As always, forewarned is forearmed.
Creating a Table via DDL More specifically, you'd describe the names and descriptions (data type, size, etc.) of each of the columns in the table. For instance, you might enter the following command:
CREATE TABLE Tasks ( Don't let some of the options throw you. It's really pretty simple. This creates a table called Tasks with several columns. The first, TaskID, is defined as a COUNTER field. That's the ODBC term for what's called an AutoNumber field in Access and an Identity Column in SQL Server. You don't use Auto-Number when using DDL to configure or reconfigure columns for an Access table. You have to use the proper SQL "data type." Notice the other types of columns created, including TEXT, SHORT, MEMO and BIT. Table 1 shows several of the available data types as defined in the Access "Jet Engine" SQL reference. Some of them will work in SQL Server and other DBMSs as well. See the documentation for your specific database for more details. (If any aspect of this DDL fails in a database other than Access - toward which this article is primarily focused - see your database documentation to find the equivalent statements.) Note too that the TaskId column was defined as the primary key. You can do that via DDL as well. We even defined the Requested_By column as a foreign key with a CONSTRAINT clause defining a relationship to the Employees table, specifically the EmployeeId column. Finally, note that you can also indicate if a column is allowed to take on NULL values by adding the NULL keyword.
Entering the Commands via CFQUERY There are two potential traps here: 1. The Datasource definition in the CF Administrator may be set to prevent any SQL statements other than SELECT, INSERT, UPDATE, DELETE and/or stored procedures. This is an option available under the "CF Settings" button for a given data source in the Administrator. Be sure that all five checkboxes are turned off. 2. Be sure that you have the permissions to enter these commands. Sadly, in most Access databases there's no security defined to prevent this, but if there is, or if you're using an enterprise-class database, you may need to specify an authorized user ID in either the CFQUERY (via the USERNAME and PASSWORD attributes) or in the Datasource definition in the Administrator. Again, see the "CF Settings" button there.
Performing Additional Reconfiguration via DDL Let's say you want to add an index to an existing table on your remote database. It's as simple as:
CREATE INDEX Completed ON Tasks
Or to add a relationship definition:
ALTER TABLE Employees Note that when defining relationships in both the ALTER TABLE and CREATE TABLE, we can't indicate cascading deletes or updates (in Access, at least). We also can't modify relationships. We must drop and re-create them. That's easy, too. It's our first look at the DROP statement:
ALTER TABLE Employees
ALTER TABLE Tasks
Dropping is just as easy:
ALTER TABLE Tasks (Although SQL supports a RE-NAME and MODIFY clause for the ALTER TABLE statement, in Access - the Jet Engine - once it's created you can't alter the data type of a field or change its name. The only way to convert an existing field from one data type to another is to add a new field, run an update query to populate the field with values from the original field, then drop the original field.) You can also delete or drop an entire table with: DROP TABLE Tasks This is a dangerous command, of course. Use it with caution! (One potential source of confusion: notice that you don't drop columns or indexes via a DROP statement. Instead, you drop them by way of a DROP clause within the ALTER TABLE statement, a subtle but important difference.)
Some Closing Thoughts One other potential gotcha: if any of your column or table names are reserved words in SQL, you can surround them in square brackets. That protects you from SQL errors if you happen to use reserved words for table or column names. Of course, you may omit the brackets and let the error occur during this process to catch you making the mistake. But sometimes the failure in the SQL won't be so obvious. The choice is yours.
For More Information These books are also available online at Microsoft, in their MSDN library. Visit the following link to see the data types, http://msdn.micro-soft.com/library/sdkdoc/daosdk/dajsql05_1yr7.htm. From there, use the "Show TOC" link. Look for the specific DDL statements to be covered (strangely enough) under "Microsoft Jet Database Engine SQL Reserved Words," which appears a few sections down in the Table of Contents. Both the Access Help file and Web site also offer examples. In addition, there's an interesting discussion of some of the limitations of doing SQL in the Jet Engine at http://msdn.microsoft.com/library/psdk/dasdk/odbc50qb.htm. Finally, the Web site also offers an online book, the "MS Jet DB Engine Programmer's Guide." And Chapter 3 of that book, available at http://msdn.microsoft.com/library/books/dnjet/c3_body.htm, offers an even better written discussion of DDL topics. Be sure to open the TOC and then the subsections of the chapter to see those that are headed "...by Using SQL DDL," such as "Creating and Deleting Indexes by Using SQL DDL." It covers creating and deleting tables, columns and relationships as well.
CFDJ LATEST STORIES . . .
SUBSCRIBE TO THE WORLD'S MOST POWERFUL NEWSLETTERS SUBSCRIBE TO OUR RSS FEEDS & GET YOUR SYS-CON NEWS LIVE!
|
SYS-CON FEATURED WHITEPAPERS MOST READ THIS WEEK |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||