Transcoding

We refer to "Transcoding" the need to match the codes coming from two different systems. An entity will for example be uniquely identified by one or more codes in a first system (eg name, surname and birthday) and other identifiers in another system (eg no social security). To interconnect the entity (eg Person) in both systems it is necessary to establish a correspondence codifications table.

Issue

The goal is to ensure that the correspondence table is current and complete. Points to consider include the following:

• Get the new codes of the two systems as soon as possible,
• Identify non transcoded codes of each systems, "isolated"
• Make sure there are no duplicates, ie, one code of the system 1 mapped to two or more codes of system 2 and vice versa, "duplicate"
• Provide a management interface with filters and sorts to group matches, isolated and replicated,
• And possibly other functions to increase efficiency:
  
  • Alert the administrator of correspondence (mail with the isolated and duplicate list)
  • Establish an automatic processing of correspondence generation (if a rule can be programmed).

Using Data Studio for transcoding

Use of Data Studio offers several advantages:

• Direct access to two data sources,
• Pre-treatments for transcoding:
  
  • Extract the new keys,
  • Dedupe keys,
  • Alert the administrator of the arrival of new keys.
• Using DataStudio Matching module:
  
  • Graphic implementation of matching processes,
  • Matching of N columns with M columns,
  • Identification of the isolated key on one and the other side,
  • Identification of duplicate keys inside the transcoding table (not bijective)
  • Definition of the precision for floats and dates,
  • treatment of significant volumes without saturation of machine resources,
  • Automatic generation of the matching results tables with key originals columns name for use in downstream processing,
  • Presentation of results in a dedicated interface,
  • Consultation and maintenance of matchings graphically
• Post-taitements to transcodification:
  
  • Automatically generate transcodifications (if available algorithm)
  • Alert administrator isolated keys and duplicate keys.

Example transcodification

Graphic setting the left side transcodification

Graphic setting the right side transcodification
 

Automatic generation of the results tables with the originals columns name that could be directly managed into interfaces to be developed

Track Results Matching in DataStudio interfaces

Example interface to manage transcoding