I am hoping to use the Solr QueryElevationComponent for best bets on my site search. Hooking it up was easy enough, the biggest issue I have is _uniqueid field which is used by the component to identify the document (because it's defined as the uniqueKey in the schema). The _uniqueid field by default contains the index name and the version number which makes using the component quite unmanageable.

eg.

<query text="boo">
    <doc id="sitecore://master/{0a5cce11-62fa-40da-ad28-6ec05e236b0b}?lang=en&amp;ver=18&amp;ndx=sitecore_master_index" />
</query>

Has anyone got any ideas on how I can get this to work? I have thought of a few options, but have no idea on the repercussions or if they are even possible:

• Rewrite the _uniqueid to something more usable and static between versions
• Use a different field as the uniquekey (e.g. _group)
• Store only the latest version of the Sitecore item in Solr so I can use the GUID as an identifier.
• List the next few document versions (18,19,20 etc) in the xml file in current format to handle version updates to the Sitecore item