Number of items in the Review set for a Discovery Accelerator Case does not match the number of items Exported from that Case
The total number of items in the Review Set for a Discovery Accelerator (DA) Case is greater than the total number of items in a DA Export of all items from that Case. All Searches in the Case were run with the 'Include items already in review' option de-selected and the 'Automatically accept' option also de-selected. All Searches were manually initiated (ie. no Searches were scheduled Searches) and manually Accepted.
For example, the Review Set shows 5000 items captured by all Searches; an Export of all items from that Search shows only 4500 items.
This behavior is by design and can be explained as below.
An item is entered into the Review Set one time within the boundary of a Case based on their Enterprise Vault SavesetID by the first Search that was Accepted that contained the item in the Search results. A unique ID, called the DiscoveredItemID, is created for the mapping between the Enterprise Vault SavesetID and the DA CaseID. The DiscoveredItemID is also cross-referenced to the Search that found it. All of these references are stored in separate Tables within the DA Customer Database on the SQL Server. Accepting a Search causes the items to be entered into the Review Set. A Search that has completed but has not been Accepted, ie., is in a Pending Acceptance status, will not have its results placed into the Review Set until the Search has been Accepted.
It is important to understand that the sequence of initiating and Accepting Searches can have a bearing on the number of items in the Review Set of a Case and the number of items available for Export from that Case.
The following examples clarify this behaviour, which is expected and by design:
Example 1 - Searches initiated and Accepted before initiating more Searches
A Search was initiated, allowed to complete and Accepted, another Search was initiated, allowed to complete and Accepted, then another Search was initiated, allowed to complete and Accepted. All Searches have the 'Include items already in review' and 'Automatically accept' options de-selected. The expected behavior is that the number of items found by the Searches should match the number of items available for Export. If the first Search found 3000 items, there would be 3000 items added to the Review Set for that Case. If the second Search found 1500 items, there would be an additional 1500 items added to the Review Set for that Case. If the third Search found 500 items, there would be an additional 500 items added to the Review Set, for a total of 5000 items for that Case. All of these items would be unique as the 'Include items already in review' option was not selected, causing each Search upon completion to compare its results with the items in the Review Set and discard any items that already exist in the Review Set. An Export of all items would result in 5000 items exported.
Example 2 - Searches initiated and Accepted after initiating more Searches
If a Search was initiated, then another Search was initiated, then a third Search was initiated, but none of the Searches were Accepted, the Acceptance process might lead to a different outcome, as are no items in the Review Set with which to compare the Search results. Using the same result counts as above, assume the first Search found 3000 items, the second Search found 1500 items and the third Search found 500 items. If the second Search with 1500 items was the first to be Accepted, its 1500 items would be added to the Review Set for that Case. If the first Search with its 3000 items was then Accepted after the second Search had completed the Acceptance process items and its items were found to be unique, ie. not present in the Review Set for that Case, then its 3000 items would be added to the Review Set. The total count in the Review Set is now at 4500. If the third Search was now Accepted and it was found that its results are not unique, ie., did exist in the Review Set (as there are now items to be compared against), then it is entirely possible no items from this Search would be added to the Review Set as they already exist, but the items from this third Search would still be available for Review. This is due to the cross-reference created between the item found in the Search and the Search itself within the Tables in the Customer Database, as previously explained. As a result, the Review Set for this Case would show 5000 items, but the Export would only show 4500 items.
A method to alleviate this scenario is to allow each Search to complete and to be Accepted before initiating the next Search. However, as this is not always feasible, the examples above should demonstrate the behavior seen and also explain the expected nature of the behavior.
It is important to note that in these examples there are no differences in the results found. The only difference is in the Export due to the sequence of Accepting the Searches.
The first 2 examples are by design, as explained in the examples above.
Here is an example of behaviour that may need to be investigated: Indexes report different numbers of items than Storage actually contain. For example, the Search may report 5000 items found, and an Export of that Search completes with 4500 items Exported. A second Search run at a later date using the same criteria returns the expected 4500 hits and an Export of the second Search completes and Exports the same 4500 items. This scenario indicates a possible miss-match of Index items counts and Storage items counts and should be investigated from an Indexing perspective.
Was this content helpful?
Rating submitted. Please provide additional feedback (optional):