TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Data Profiling Pros Welcome TS Discovery 5.0

Drums keep pounding rhythm to the brain—and the data quality beat keeps on keeping on. Case in point: Trillium Software last week announced a new version of its TS Discovery 5.0 data profiling tool. Analysts are calling it Trillium’s strongest profiling offering to date.

• By Stephen Swoyer
• January 11, 2006

Drums keep pounding rhythm to the brain—and the data quality beat keeps on keeping on. This in spite of the clouds that have, of late, lour’d over (not just the House of York) but the data quality marketplace as a whole.

Last week, for example, Trillium Software (a division of Harte Hanks Inc.) announced a new version of its TS Discovery 5.0 data profiling tool.

Analysts are calling the new TS Discovery 5.0 Trillium’s strongest profiling offering to date. And in a marketplace that’s still trying to find its feet in the aftermath of Pitney Bowes Inc.’s abortive acquisition of data quality pure play Firstlogic Corp., that’s not a bad thing.

Trillium acquired its data profiling technology two years ago from the former Avelino Technologies, and has since augmented it with a pair of new releases, first in late 2004 (TS Discovery 4.0) and then in the first quarter of last year (TS Discovery 4.1, which delivered UniCode support). But neither of these releases can hold a candle to TS Discovery 5.0, says Rob Lerner, a senior analyst for data management with consultancy Current Analysis Inc.

“Certainly, this version of TS Discovery is the strongest version yet of the data profiling solution,” he writes, citing tightened integration with Trillium’s bread and butter data quality technology as a particularly strong selling point. “Trillium Software has been working on the integration of the two products almost from the beginning, before the acquisition of Avellino, and each succeeding release offers additional improvements in this area.”

For example, Lerner says, TS Quality users can now launch TS Discovery functions without leaving the former program. Moreover, TS Discovery ships with data cleansing capabilities, which can be executed from the TS Discovery GUI. On top of this, users can amend or recode data structures and values from within TS Discovery; re-profile cleansed data; and export cleansed and profiled data in TS Discovery to files.

“Moreover, TS Discovery now enables users to run business rules against any source system. Prior versions enabled users to build and share business rules, but in 5.0 users can actually externalize the execution of the rules against any data source,” Lerner says. “This expands the capabilities of the solution, and effectively enables one to use the capabilities to monitor data and business rules standards at their sources.”

Other improvements include support for selective profiling of sources, i.e., users can configure TS Discovery to profile only specific rows and columns. What’s more, Lerner notes, users have the option of doing so in the source system itself. “This speeds processing and enables the user to profile only the data that’s important to [them] or to launch single queries in a particular system, eliminating unnecessary profiling,” he points out.

Elsewhere, the reloaded TS Discovery 5.0 incorporates GUI enhancements, including GUI-based support for data manipulation and modification (e.g., users can create new files in both relational and non-relational data by creating joins, or by adding new attributes based on business rules).

Nevertheless, Lerner says, Trillium’s work isn’t quite done. “[T]he data quality capabilities in TS Discovery are limited to name-and-address data, although TS Discovery can profile non-name-and-address data. There are also limitations with respect to the information that can be appended to data handled by the product,” he observes. “Finally, among the range of countries currently supported by TS Discovery, few require Unicode and some of the countries supported by TS Quality are not among those included.”