Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.


This is now in DMTN-153


The DPDD (LSE-163) represents an "idealized" data model that cannot directly be used as a recipe for the construction of a physical database schema and is not a full, precise definition of a data model.  It should be treated as a requirements document for the data model and schema rather than itself defining them.


As will be seen elsewhere in this document, we expect to link the DPDD with additional metadata defining the physical data model for the released data products.  However, to facilitate configuration management, the DPDD source document will contain no back-links to the data model definition, so that the DPDD repo in Github can be placed under configuration control on its own.  We expect the DPDD itself to be changed infrequently, and to continue to be under the higher, project-wide level of change control that it currently is, while the associated data model may evolve more often as construction proceeds and algorithmic choices are further resolved.

The Science Data Model

The Science Data Model (SDM) specification is a machine-readable specification for the physical data model for the publicly released science data products of the LSST Project.


The SDM-to-DPDD mapping information may be useful as part of the documentation of the Science Data Model that LSST exposes to users, to facilitate their understanding of how the SDM corresponds to the DPDD.  To this end, for instance, it may be appropriate to include in the TAP_SCHEMA tables an additional column (something permitted by the standard) that provides, where appropriate, the DPDD Identifier for a data item.

Mapping from Science Pipelines outputs to the Science Data Model - "Standardization"

The SDM specification does not directly constrain the data models produced by the Science Pipelines code's Tasks and PipelineTasks.  In some cases, for instance, the Science Data Model will envision a table composed of information assembled from the outputs of several pipeline stages.


Colin Slater and Yusra AlSayyad have in recent months been working on prototyping elements of the above system.

Data model metadata for Science Pipelines outputs

The SDM specification will fully enumerate the metadata that LSST provides to its users describing the Science Data Model (e.g., its UCDs).  This metadata is used, in part, to enable intelligent behavior in the Science Platform components when serving data to users.