On 7/9/2013 3:38 PM, Katie McCorkell wrote:
I am curious about the "Deleted Docs:" statistic on the solr/#/collection1
Overview page. Does Solr remove docs while indexing? I thought it only did
that when Optimizing, however my instance had 726 Deleted Docs, but then
after adding some documents that number decreased, eventually to 18 Deleted
Docs.
I understood these Deleted Docs are from situations where two docs have the
same UniqueKey. However my data had way more deleted docs than I expected.
I was using a data-generated uniquekey, when I changed to using the UUID
generator there were 0 deleted docs. But I just wanted to double check, are
there any other cases which would create a Deleted Doc?
Changes to deleted documents can happen through normal segment merging.
Optimizing is just an explicit and deliberate merge down to a single
segment, but segment merging is a normal part of Solr/Lucene indexing.
Any deleted documents in segments that get merged will be purged.
I believe the UUID generator will always generate a new value even if a
document with the same information in the other fields is indexed again.
This option should only be used if you do not have an existing field
with unique values on every document.
Thanks,
Shawn
