[
https://issues.apache.org/jira/browse/NUTCH-963?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13008421#comment-13008421
]
Markus Jelsma commented on NUTCH-963:
-------------------------------------
Solr deduplication makes its own (fuzzy) hashes on one or more fields. Separate
algorithms on different fields can be combined. It does not take into account
the score of a document if you mean the index-time boost on the document. But
if there is a separate score (or boost) field then a combined signature on
body, title and boost will work.
All aside, i agree we should go for a single Nutch command for cleaning an
index, doing dedup and/or 404 cleaning in one swift go.
I'll rereview this patch and do further testing and won't forget CHANGES.txt.
After that i believe we can create a new related issue for the new
deduplication.
> Add support for deleting Solr documents with STATUS_DB_GONE in CrawlDB (404
> urls)
> ---------------------------------------------------------------------------------
>
> Key: NUTCH-963
> URL: https://issues.apache.org/jira/browse/NUTCH-963
> Project: Nutch
> Issue Type: New Feature
> Components: indexer
> Affects Versions: 2.0
> Reporter: Claudio Martella
> Assignee: Markus Jelsma
> Priority: Minor
> Fix For: 1.3, 2.0
>
> Attachments: NUTCH-963-command-and-log4j.patch, Solr404Deleter.java,
> SolrClean.java
>
>
> When issuing recrawls it can happen that certain urls have expired (i.e. URLs
> that don't exist anymore and return 404).
> This patch creates a new command in the indexer that scans the crawldb
> looking for these urls and issues delete commands to SOLR.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
