About Importing and Exporting Data


Hyperstage provides three ways to import data:

INSERT is described in Data Manipulation Statements and is the slowest load approach. The DLP is the fastest load method but supports less load syntax.

Distributed Load Processor (DLP)

  • Fastest loader
  • Less error handling diagnostics (only the source file row number pertaining to the error is returned)
  • Strict input file formats (supports delimited text and binary formats)
  • Variable Data Pack size
  • Load-time clustering


  • Supported by virtually all ETL tools
  • Can be very slow depending upon the approach and commit rate

If you are using an ETL tool, using the DLP or COPY TO method with the binary format is most efficient, although this approach may require more data preparation. For large fact tables, using the DLP or COPY TO method with either binary or text input is recommended.


  • Supports capability similar to the DLP (does not support cluster on load or varying packrow size in this release)
  • Primarily used for local instance only
  • File-based loading
  • Can work within transactions