It all depends on how you want to proceed -- You can use an ArcPy script to make a
Query Layer request and extract the data from the source database, then insert it
into a truncated target table (after the service which would lock the truncate has
been stopped). Or you can link the servers and query the export table directly
from the ArcSDE connection. Or you can use Well-Known Text in an ASCII export
to repopulate the table with a command-line utility like 'asc2sde'.
Detecting change over time isn't all that difficult if you have access to a cryptographic
hashing function to generate {key,hash} pairs (based on rowid column, usually).
Two skinny tables and three views can be used to identify the insert, update, and
delete events. You can even run the analysis in realtime to update the content
more frequiently than daily, and it wouldn't require any downtime. All you need is
a location for the intermediary files.
I must point out that proper tuning of a Geometry layer could produce better
performance than a SDEBINARY layer, so you might not be asking the right
question.
- V
