Different versions of ArcMap creating mosaic datasets in Oracle

If different versions of ArcMap are connected to the same Oracle database, do they behave the same as "Create Mosaic datasets"?

The answer is: not the same, there will be subtle differences

In this example, the version of ArcMap is 10.2.2 with 10.3, and the version of the database server is 11.2.0.3,SDE Repository version is 10.1.

1. Create a mosaic dataset using the 10.3 version of ArcMap MD1003

   Before creating a mosaic dataset, create a user_objects backup table user_objects_b;

   After the mosaic dataset is created, the database adds the following tables, sequences, and triggers.

   Added 1 triggers, 10 tables, 7 sequences.

   Here are two points of knowledge, I'll introduce you a little bit.

   First, what does the 10 in raster_layer_10 and sde_ras_10 name mean?

   This value actually represents the Amd_md1003_cat in SDE. The value of the rastercolumn_id field in the Raster_columns table, such as:

   Second, what does the number of r120-r124 these series represent?

   This value actually represents the value of the registration_id field in the Sde.table_registry table for 5 tables starting with AMD, such as:

   R120-r124 these sequences exist to assign values to the Objectid fields of the corresponding table.

   So how are these sequences defined? Let's take a look at:

   Note that the maximum value for this sequence is 1028.

2. Create a mosaic dataset using the 10.2.2 version of ArcMap MD1002

   Delete the user_objects_b and recreate the user_objects backup table user_objects_b;

   After the mosaic dataset is created, the database adds the following tables, sequences, and triggers.

   Added 1 triggers, 9 tables, 6 sequences. Compared to creating a mosaic dataset with version 10.3, a amd_%_def table and a sequence corresponding to that table are missing.

   At the same time, it is known that the rastercolumn_id value of Amd_md1002_cat is 11, such as:

   The sequence starting with R corresponds to the AMD opening table as follows:

   The sequence is defined as follows:

   Note that the maximum value for this sequence is 2147483647, which is much smaller than the maximum number of sequences created by version 10.3.

   Conclusion:

   1. For the Create Mosaic dataset operation, although the database version and the SDE repository version are consistent, the differences in the ArcMap version still lead to different recursive operations. This can be inferred: for other operations, do not exclude due to different versions of the resulting behavior of different situations occur, so the test and the final production process, it is best to use the same version of the ArcMap version.
   2. A mosaic dataset created with the 10.2.2 version, with a sequence maximum of 1 billion levels, and a value of 16 per increment, so that the available values are only billion, and for a large mosaic dataset (containing multiple datasets, or multiple delete re-entry scenarios), it is very likely that the sequence will be filled with the added data set failing. (also, because the cache value of the sequence is 20, if the instance is closed frequently, or if the "alter system flush Shared_pool" is often executed, it causes the sequence to produce many "holes" that speed up the maximum limit of the breakout sequence)

Different versions of ArcMap creating mosaic datasets in Oracle