I suffer sometimes from data loss after compressing ArcSDE geodatabase, this geodatabase contains versioned feature classes, so the compressing script reconciles and compress the geodatabase:

# Execute the ReconcileVersions tool.
arcpy.ReconcileVersions_management(db_path, "ALL_VERSIONS", "sde.DEFAULT", versionList, "LOCK_ACQUIRED", "NO_ABORT", "BY_OBJECT", "FAVOR_TARGET_VERSION", "POST", "DELETE_VERSION", "c:/reconcilelog.txt")

# Run the compress tool.

it seems that the main reason is a bad use of ReconcileVersions_management, what are the parameters recommanded ? and how i can test if data loss would happen again ?

the geodatabase is based on ArcSDE 10.1 and Postgre SQL 9.0.5.

  • Please define "accidental data loss" -- what exactly is happening? – Vince Dec 12 '13 at 17:22
  • i make edits normally , so when the database gets slowing, i run the compacting script, sometimes i i can have nothing lost, but sometimes i can have some data loss, i've tried to find a logical reason , but i have found nothing. – geogeek Dec 12 '13 at 17:25
  • 1
    What does "data loss" mean to you? Note that compressing a versioned geodatabase is not the same as compacting a file geodatabase. – Vince Dec 12 '13 at 17:32
  • 1
    With batch reconcile it is quite possible to ask to have your data discarded. While it might qualify as "accidental data loss", it's not a result of software failure or server performance. – Vince Dec 12 '13 at 18:23
  • 5
    I'm not familiar with the ArcPy tool, but if you say NO_ABORT and FAVOR_TARGET_VERSION (in your case 'DEFAULT')...I think you have effectively instructed the tool to "ignore the conflicts" and take what is in DEFAULT. So you would "lose" data whenever a conflict occurs. Additionally, you defined your "conflict" as "BY_OBJECT" so even if the same feature had different modified attributes...it's still a conflict. If you go "BY_ATTRIBUTE", you are only conflicted when the same attribute is modified. – Jay Cummins Dec 12 '13 at 18:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.