Transcend is comprised of two distinct yet integrated products: Transcend ETL and Transcend Files. Each product addresses a void in all the ETL tools currently on the market.

Transcend ETL

Batch load processes for enterprise data warehouses require loading millions or sometimes billions of rows daily, and issuing standard update and insert statements in isolation is often insufficient for these data volumes. Furthermore, Oracle Database Server contains features for improving these load routines, either by making them run faster, by improving availability of the data warehouse while they are running, or both.

Constraint and Index Maintenance

Transcend ETL allows the disabling and enabling of constraints based on numerous factors: table owner and name, constraint type, regular expression, etc. Transcend ETL is also capable of disabling constraints without specifying the table the constraint is on, but instead, by specifying the table the constraint references.

Index maintenance is equally powerful and efficient. Again, you are able to specify things like table owner, table name, index type, regular expression, etc. But Transcend ETL has some additional benefits when dealing with partitioned indexes. When loading data in a large, partitioned table, such as a fact table, you are usually only interested in marking indexes unusable that correspond to the data set being loaded for that day. For instance, when loading a day's worth of data, there is no reason to mark indexes unusable on partitions for last month or last year. Transcend ETL allows you to only alter the indexes affected by the data set being loaded.

Rebuilding indexes is also easy. When specified, Transcend ETL uses the Oracle Scheduler to execute the rebuild processes in the background so that multiple statements are executed concurrently. This drastically reduces index rebuild times, and this functionality is also available for other index and constraint-related processes. Specifying the concurrent parameter in any index or constraint management task will cause the process to run in the background, where Transcend ETL will coordinate the completion of the set.

General Object Management

Transcend ETL provides object management for data warehouse ETL cycles and large OLTP batch load operations. In addition to altering and rebuilding indexes in the examples above, you will be able to easily drop multiple objects--such as indexes and constraints--with a single call, as well as build tables, indexes and constraints based on objects already existing in the database.

Complex Load Routines

Transcend ETL facilitates complex data warehouse and batch load procedures using a mix of configuration tables and singular procedure calls.

• Multiple object management tasks, including index maintenance, constraint maintenance, partition-exchanges, etc., can be grouped together in objects called mappings.
• Multiple tables matching a similar naming standard can be loaded from one schema--either local or remote--to another, with a single procedure call.
• Performing partition-exchange load scenarios, including managing all indexes and constraints.
• Loading dimension tables containing a combination of Type 1 and Type 2 SCD columns using complex SQL analytics to perform the entire dimensional load with a single statement.

Transcend Files

Mainframes and legacy systems are still prevalent in enterprises today, but new file-based integration standards such as XML are increasing the role of files in non-legacy systems as well. But for some reason, none of the ETL tools take serious the role that files play in the data warehouse space. Not only do we need the ability to move files around--which most the ETL vendors do support--but we also need the ability to catalog all the files received, including the archival of files in case they are needed later. But we also need to know all of the following: when was the file loaded, how many rows did it contain, were all of those rows loaded successfully? Just as important, we need to assign business logic to our load processes to determine whether to load a particular file at all. And we should be able to do all of this without writing a single line of code.

One-Step Configuration of Feeds and Extracts

A simple configuration process tells Transcend Files everything it needs to know about file feeds: what sort of name to expect, where to look for them, where to put them, and acceptable file sizes. But equally important is letting Transcend know what to do in case multiple files are found matching the naming standard: should the oldest or newest version be loaded? Perhaps all versions of the file should be loaded, or maybe multiple files is a sign of an error, and the process should simply abort.

Configuring extracts is just as easy. Transcend simply needs to know what object represents the data to be extracted: either a table or view. You also configure the location to deliver the file and where to archive it. Finally, let Transcend know how to present the file. Should column names be included as headers? What is the date format desired? What delimiter do you prefer and should quotes be placed around the column names?

Full Auditing and Notification

Transcend stores full auditing about all your feeds and extracts. But also, you are capable of configuring notifications and notification lists for information about your files. Maybe you want to know if the number of rows rejected for a particular file exceeds a predefined threshold. Or maybe you want a copy of a particular file emailed to a distribution list. Transcend Files can handle all of this for you, and once again, without a line of code.

For more information about Transcend or our other product offerings, contact us here.