[
https://issues.apache.org/jira/browse/SOLR-1483?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12762181#action_12762181
]
Mark Miller commented on SOLR-1483:
-----------------------------------
Attaching that javadoc from Lucene:
* <p>Good values for <code>precisionStep</code> are depending on usage and
data type:
* <ul>
* <li>The default for all data types is <b>4</b>, which is used, when no
<code>precisionStep</code> is given.
* <li>Ideal value in most cases for <em>64 bit</em> data types <em>(long,
double)</em> is <b>6</b> or <b>8</b>.
* <li>Ideal value in most cases for <em>32 bit</em> data types <em>(int,
float)</em> is <b>4</b>.
* <li>For low cardinality fields larger precision steps are good. If the
cardinality is < 100, it is
* fair to use {...@link Integer#MAX_VALUE} (see below).
> Example schema is confusing with int, tint and pint fields
> ----------------------------------------------------------
>
> Key: SOLR-1483
> URL: https://issues.apache.org/jira/browse/SOLR-1483
> Project: Solr
> Issue Type: Bug
> Components: documentation
> Reporter: Shalin Shekhar Mangar
> Fix For: 1.4
>
>
> Example schema defines int (TrieIntField), tint (TrieIntField) and pint
> (IntField) which is confusing. In particular, the comments for int fields
> tell users to use tint types (which is the same thing).
> Let us remove tint, tlong, tdouble, tdate types from example schemas and fix
> the comments - carefully noting the things trie fields do not work with.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
