Deduplication of external organizationsDeduplication of external organizations

This article details the deduplication of external organizations according to the automatic process.

Identification and validation

The following chart shows how the system identifies and validates merge candidates:

Merge target identification

If one of the items has a SciVal ID, then that item is the target of the merge. If the items were identified based on a classified ID other than SciVal, or if none of the items has a classified ID, then the oldest item by date created is the target of the merge.

Merge

The system merges values in the following fields according to the default merge logic:

• TYPE_CLASSIFICATION
• NATURE_TYPES
• NAME
• ACRONYM
• ALTERNATIVE_NAMES
• IDS
• ADDRESS1
• ADDRESS2
• ADDRESS3
• ADDRESS4
• ADDRESS5
• POSTALCODE
• CITY
• STATE
• COUNTRY
• SUBDIVISION
• PHONE
• MOBILEPHONE
• FAX
• EMAIL
• BANKACCOUNT
• VAT
• LINKS
• KEYWORD_GROUPS
• NOTES
• WORKFLOW

The system applies specific logic for values in the following fields:

• Hierarchy: In addition to the default logic, the following additional logic applies: If there is a parent organization on the source but not on the target, and that parent is not the same organization as the target organization itself, then the system adds the parent to the target. If the source is the same as the parent on the target, then the system removes the parent. If there is a parent on the target but not on the source, then the system keeps the parent on the target.
• Taken over by: The system does not apply the default logic. Instead, the the system cleans up the values as follows: If the field is defined on the target then the system iterates over the values. If the items evaluated do not exist on the source, or are equal to the source, then the system discards them.
• Visibility: If the visibility of the target is less restrictive than the visibility of the source, then the system sets the visibility of the target to the visibility of the source. If the source is marked as confidential and the target is not, then the system sets the visibility of the target to confidential.