4 Replies Latest reply on Aug 7, 2020 7:37 AM by Ashley DCruz

    EDC & GIS

    Ashley DCruz New Member

      Does anyone have any experience with documenting GIS databases within EDC?

        • 1. Re: EDC & GIS
          Venkatesh Srinivasan Active Member

          Hope you are talking about ArchGIS or ESRI.

          Currently, EDC does not have a native scanner for GIS systems like ArchGIS or ESRI.

          I see an Enhancement request (EIC-43621) to add support for this GIS scanners in one of the future releases.


          If you can convert the data from ESRI to a csv or XML file, it can be ingested to EDC using a File System scanner.


          If you have the GIS data (like latitude or longitude) present in any supported source systems, you can scan it through the native scanners and classify them using data domain discovery or business Titles or custom attribute tags.



          • 2. Re: EDC & GIS
            Ashley DCruz New Member

            Hi Venkatesh,


            Thanks for your response. Yes I am referring to ESRI. The users particular question was if 'feature data sets' were displayed in the catalog. They are all have something in common like projection, how they are edited and how they are used.


            Our GIS database is in Oracle and we have ingested that but we do not see the feature data sets.


            I am not a GIS person so hopefully my question makes sense.




            • 3. Re: EDC & GIS
              Srinivas Pai Guru

              Hi Ashley


              Could you please explain how did you implement the feature datasets on your Oracle database?




              • 4. Re: EDC & GIS
                Ashley DCruz New Member

                Hi Srinivas,


                Here is an example


                The items 'TRANS.Transmission, Trans.RiskDataSet, Trans.Uptime' are called 'Feature Data Sets' in GIS.

                They are a collection of tables that have something in common – like Projection, how they are edited, how they are used.  Here's an example of the GISTP.Transmission Feature Dataset. There are probable 30 or 40 more feature Classes (tables) in the Transmission Feature Dataset.

                The event tables are actual tables, and are not part of any Feature Dataset




                Versioned Tables

                Almost all tables and Feature Classes in our GIS database are versioned tables.

                This means that the database usually has a different count for the base table than what is actually is the correct amount of records.

                For example, we  have a table in GISP and GISV named 'AGL.GASMAIN'  When you query it, you get  a  different value than you get when you use the multiversioned view.


                select (select count(*) from agl.gasmain) "AGL.GASMAIN", (select count(*) from agl.mv_gasmain) "SDE.DEFAULT" from dual;

                1. AGL.GASMAIN SDE.DEFAULT

                ----------- -----------

                790789 799644


                Note the counts are different.  The 799644 is the correct value of active GasMain segments in the GISP database.

                When you use ESRI products, you don't have to take this into consideration, as it knows the DB is a versioned Database and used the Adds and Deletes table to give you the correct answer of 799644.


                For AGL.GASMAIN, there are 2 additional tables when versioned, AGL.A62 and AGL.D62 (Adds and Deletes).

                These are used in a complex join statement to return the correct values for the GasMain table at a particular state.

                When you Add a record, the StateID and values are actually placed in the Adds table (A62)

                When you delete a record, the StateID and OBJECTID is placed in  a row in the Deletes table (D62)

                So, the values in the AGL.GasMain table at the SDE.DEFAULT is

                BaseTable + (Adds at the SDE.DEFAULT State) – (Deletes at the SDE.DEFAULT State)


                Currently, not all of our versioned feature Classes / tables have views created; but  a vast majority of them do.


                So when informatica ingests the table it should only really ingest the latest view and not all the underlying tables which would need to be imported.