[ 
https://issues.apache.org/jira/browse/NUTCH-1867?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14159693#comment-14159693
 ] 

Sebastian Nagel commented on NUTCH-1867:
----------------------------------------

> do we need to state on the conf property description that this is of type 
> float? or is this not required?
The property is only used to pass the min score to the mapper. It is not listed 
(and documented) in nutch-default.xml because setting it in a config file is 
useless, it is always overwritten either by the command-line value or by 0.0

> CrawlDbReader: use setFloat to pass min score
> ---------------------------------------------
>
>                 Key: NUTCH-1867
>                 URL: https://issues.apache.org/jira/browse/NUTCH-1867
>             Project: Nutch
>          Issue Type: Improvement
>          Components: crawldb
>    Affects Versions: 1.9
>            Reporter: Sebastian Nagel
>            Priority: Trivial
>             Fix For: 1.10
>
>         Attachments: NUTCH-1867-v1.patch
>
>
> The float value "min" score in the CrawlDbTopNMapper is passed via property 
> "db.reader.topn.min" as a long (multiplied by 1Mio.). The comment "no 
> setFloat() in the API" is no longer valid, the method exists in 
> [Configuration|https://hadoop.apache.org/docs/current/api/org/apache/hadoop/conf/Configuration.html|Configuration]
>  and should be used. Reported by [~lewismc], see NUTCH-1857.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to