I just stumbled upon an interesting behavior:
- Have a document set:
- Title: My Special Document Set
- Perform a crawl (continuous / incremental crawl)
- Search returns 1 result.
- Rename the document set with different casing: "my special docuMENT SET"
- Perform a continuous / incremental crawl
- Search returns 2 results. One for each case (my special... & My Special...)
DocIdmanaged property is different for both search results
- The Duplicate Rows are 0 or in other words: Total Rows: 2, Total Rows including Duplicates: 2, so
trimduplicates=truedoes not help, still 2 results.
- When I now perform a full crawl, the duplicate entry is removed.
It seems as continuous and incremental crawl do not take care of casing/renames of folders/FileLeafRef and create a duplicate entry for when only the casing of an item is changed. Besides running a full crawl, is there anything else I could do about this?
We're currently evaluating creating an event receiver to trigger
CrawlLog.RecrawlDocument on renames but that seems kind of non-standard and weird. The
SPListItem.UniqueItemIdis the same for the item so we could actually filter the result set. - weird that duplicate detection doesn't kick in