Elixir Repertoire for Data ETL

Debugging

• Can the procedures be run step by step? Yes
• Can the procedures be run row by row? Yes, by using a Top/Bottom N extractor.
• Can breakpoints be built into either a process or a row? Not relevant, because every procedure can run step by step at any point.
• Is it possible to define software watchpoints? Yes, by using a DataDrop processor.
• Is it possible to find and highlight any errors in the code with one mouse click? Yes, a console is provided for easy viewing of errors.

Screenshot

Functionality

• Can a data-source be read once and written to two or more different target tables/databases? Yes, by using DataDrop operator.
• As above, but then conditional on a field value e.g. Revenue > 1000 write to table 1 otherwise write to table 2 Yes, by using Subflow
• Can the rows from two different tables with the same structure be written to one table? Yes, by using to do split Concat operator.
• Is it possible to write non-normalized data structured in columns to rows? Yes, by using Invert Data operator.
• Is it possible to write non-normalized data structured in rows to columns? Yes, by using Invert Data operator.
• Is it possible to load a full table in memory and run queries on it (without joining)? Yes, with Data Cache
• Are key lookups reusable across different processes, i.e. The table is loaded once and is reusable Yes, with Data Cache
• Are slowly changing dimensions supported (hm = by hand; wizard = wizard; auto = out-of-the-box) Yes, Out-of-the-box
• Is there an ability to schedule jobs based on interdependencies? Yes, by factoring in the interdependencies in the data transforming design.
• Can the route taken by following jobs be based on errors detected in the preceding job? Yes, using OnError task in Job Scheduler.
• Is impact analysis possible for a proposed change to the database? Partial, by reading in the repository as a data source to examine dependencies. A full metadata tool is planned for further.
• Is it possible to trace the source of an attribute/data element quickly? (reversed impact analysis) Partial, by reading in the repository as a data source to examine dependencies. A full metadata tool is planned for further.
• Is it possible to publish/document a given process/transformation and navigate through it with a browser or word processor? Partial, by reading in the repository as a data source to examine dependencies and publish as a report output. A full metadata tool is planned for further.
• Is it possible to use the results of a data mining process within a given load process? Yes, through Ant Task to trigger external process before reading in the results. This is also support for ARFF, a data mining format.
• Is it possible to call analytical functions e.g. forecasting, basket analysis, regression from within the load process? Yes, through Ant Task to trigger external process before reading in the results. This is also support for ARFF, a data mining format.

Source: elixirtech.com.