Importing metadata can be a time-consuming, painstaking, and fraught task when you have multiple SAS packages to deal with. Fortunately SAS^® software provides a batch facility to import (and export) metadata by using the command line, thereby saving you from many a mouse click and potential repetitive strain injury when you use the Metadata Import Wizard. This paper aims to demystify the SAS batch import tool. It also introduces a template-driven process for programmatically building a batch file that contains the syntax needed to import metadata and then automatically executing the batch import scripts.

Read the paper (PDF) | Download the data file (ZIP)

Code review is an important tool for ensuring the quality and maintainability of an organization's ETL processes. This paper introduces a method that analyzes the metadata to assign an alert score to each job. For the jobs of each flow, the metadata is collected, giving information about the number of transformations in a job, the amount of user-written code, the number of loops, the number of empty description fields, whether naming conventions are followed, and so on. This is aggregated into complexity indicators per job. Together with other information about the jobs (e.g. the CPU-time used from logging applications) this information can be made available in SAS^® Visual Analytics, creating a dashboard that highlights areas for closer inspection.