Compliance Accelerator and Discovery Accelerator searches with specific attachment types return extra results

Problem

Enterprise Vault (EV) Compliance Accelerator (CA) and Discovery Accelerator (DA) searches that include specific attachment types return more results than expected.

Cause

One of the options available to configure a Compliance Accelerator (CA) or Discovery Accelerator (DA) search is to specify attachment types by their extension. The search behavior when specifying attachments noted in the Reviewer's documentation for CA and DA 6.0 through 10.0 SP1 states:
"If you specify one or more attachment types, only the attachments are searched, and not the items that contain them. For example, you cannot search for those items that have a specific word in their subject line or content and that contain a specific type of attachment."

This Accelerator search behavior is based on the Schema Type of the index volume(s) being searched. Enterprise Vault has two Schema Types. Schema Type 0 is 'attachment granularity' where attachments are indexed separately from the items in which they are contained. Schema Type 1 is 'item granularity' where attachments are indexed with the items in which they are contained.

User mailbox, File System, and Public Folder archives by default use Schema Type 0 until Enterprise Vault 10.0, when the 64-bit indexes were introduced with Schema Type 1. Journal archives by default use Schema Type 1. The SchemaType registry value can be used in Enterprise Vault 6.0 through 9.x to have all indexes with SchemaType 1 if desired. Note that adding this registry value only takes effect on new index volumes that are created after the registry value is added and the Enterprise Vault Indexing Service is restarted.

Solution

An Accelerator search with at least one attachment type specified that is run against an archive with Schema Type 0 index volumes will have the behavior noted in the Reviewer's documentation. Such searches will return only items that have the search term(s) within the specific attachment type(s). For example, a DA search for the term 'litigation' that has attachment types of '.doc' and '.txt' will only return items that have the work 'litigation' inside an attachment of type '.doc' or '.txt'. Any items with the term 'litigation' in the subject or content fields that do not have any attachments or do not have any '.doc' or '.txt' attachment will not be returned.

An Accelerator search with at least one attachment type specified that is run against an archive with Schema Type 1 index volumes will have the behavior of any item matching the search term(s) in the subject, content, or matching attachment types will be returned, base on the placement of the search term(s) (i.e., in the Subject or Content fields). For example, a DA search for the term 'Litigation' in the "Content only" criteria field and attachment type '.doc' will return items that have the word 'litigation' in the message body only, any item with an attachment of type '.doc' with the word 'litigation' in the attachment, or any item with the word 'litigation' in the item body and an attachment of type '.doc'. Any items without the word 'litigation' in the body or in any other type of attachment will not be returned.