[jira] [Commented] (GORA-500) bug in org.apache.gora.solr.store.SolrStore#getDatumWriter & #getDatumReader

["Kevin Ratnasekera (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/GORA-500?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16831480#comment-16831480 ] 

Kevin Ratnasekera commented on GORA-500:
----------------------------------------

{code}
  @SuppressWarnings("rawtypes")
  private SpecificDatumReader getDatumReader(Schema fieldSchema) {
    SpecificDatumReader<?> reader = readerMap.get(fieldSchema);
    if (reader == null) {
      reader = new SpecificDatumReader(fieldSchema);// ignore dirty bits
      SpecificDatumReader localReader = null;
      if ((localReader = readerMap.putIfAbsent(fieldSchema, reader)) != null) {
        reader = localReader;
      }
    }
    return reader;
  }

  @SuppressWarnings("rawtypes")
  private SpecificDatumWriter getDatumWriter(Schema fieldSchema) {
    SpecificDatumWriter writer = writerMap.get(fieldSchema);
    if (writer == null) {
      writer = new SpecificDatumWriter(fieldSchema);// ignore dirty bits
      writerMap.put(fieldSchema, writer);
    }

    return writer;
  }
{code}
Current code base no longer use schemaId to cache schema instances. Hence I am closing the issue.

> bug in org.apache.gora.solr.store.SolrStore#getDatumWriter & #getDatumReader
> ----------------------------------------------------------------------------
>
>                 Key: GORA-500
>                 URL: https://issues.apache.org/jira/browse/GORA-500
>             Project: Apache Gora
>          Issue Type: Bug
>          Components: gora-solr
>            Reporter: aakash
>            Priority: Major
>             Fix For: 1.0
>
>
> The above referenced methods use schemaId to cache DatumReader and DatumWriter instances. This restricts you to using a schema type of "map" with only a single value set. the caching method should use a more specific identifier like what is returned from fieldSchema.toString() instead of schemaId, so that the caching mechanism doesn't return a reader/writer that doesn't match your data set.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)