SE_INT64_TYPE rowids Was: Can't identify or select unversioned data in ArcMap

Question

Hi

I chanced upon this post while looking for a solution to a similar error. We have hit the 2^32-1 limit and all our feature classes has at least one column which is NUMBER(38) and that column is also the primary key.

It appears to affect ArcSDE & ArcGIS. like you stated.

I cannot change the NUMBER(38) to some thing else. A lot of procedure we have, assumes that this is a NUMBER(38). Is there any other way to get around this limitation? I can re-generate the layers, if I can somehow directly map a specific column to use SE_INT64_TYPE. Is it even possible?

Thanks for your help.

Thanks,

Alex

VinceAngelo · Answer

ArcGIS does not support an integral datatype in excess of 32 bits.  Yes, that is number one on my personal enhancement priority list (followed closely by millisecond or microsecond precision timestamps).

Unfortunately, you'll need to engineer an ID compression procedure in your context.  I recommend generating a mapping table that contains {ORIGID, NEWID}.  You should sequence and timestamp the naming of this table, so that you can preserve historic references.  After all the IDs are changed, you might find that the database takes less space (Oracle's NUMBER type is variable storage size).

- V

AlagappanNarayanan · Answer

Vince,

Many thanks for the quick response. My project schema has a number of tables and relationships, I am not sure if I can isolate those tables and do a reimport without affecting related tables.

You are correct in the sense that we don't have that many features per say, but we use the same sequencer for many other purposes and it ran past the limit now. So in essence there could be lot of gaps and holes with the range. I could compact them or resequence them but again I will have to consider the affected child tables.

Which data type will let me go beyond the INT32.

Thanks

Alex

VinceAngelo · Answer

No, you don't want to change the type of the registered rowid column -- ArcGIS Desktop will not access any table that has a column defined as SE_INT64_TYPE.  It seems unlikely that you have two-plus billion features, so the usual procedure for dealing with this limit is to export the data (to file geodatabase for maximum compatibility), then recreate an empty table or use ArcGIS to truncate the table (which should reset the objectid sequence to start at 1), then re-import the data.

If you regularly delete all features and re-add them, you should alter your procedure to truncate the table, so that the rowid doesn't grow excessively in the future.

- V