solr6.1 支持Blob字段

一种最简单的方式，就是在数据库端将Blob转换为字符:

select PROPOSAL_ID,TITLE,UTL_RAW.CAST_TO_VARCHAR2(CONTENT) as CONTENT，UTL_RAW.CAST_TO_VARCHAR2(ATTACHMENT) as ATTACHMENT from bcc_proposal

但是这种方法容易造成数据库内存问题。所以考虑另外一种方法，就是在导入solr的时候，利用BlobTransformer将Blob转换为String。继续看下文：

在上一篇文章《在Idea下编译solr 6.1源码》中，我们可以在Idea中查看solr源程序了。在本篇文章中主要介绍通过修改源码在solr中支持Blob字段的导入。
步骤一：新增BlobTransformer .java

package org.apache.solr.handler.dataimport;

import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.io.Reader;
import java.sql.Blob;
import java.sql.SQLException;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;

public class BlobTransformer extends Transformer {
    public static final String BLOB = "blob";

    public Object transformRow(Map<String, Object> aRow, Context context) {
        for (Map<String, String> map : context.getAllEntityFields()) {
            String fmt = map.get(BLOB);

            if (fmt == null) continue;

            String column = map.get(DataImporter.COLUMN);
            String srcCol = map.get(RegexTransformer.SRC_COL_NAME);

            if (srcCol == null) srcCol = column;

            Object o = aRow.get(srcCol);

            if (o instanceof List) {
                List inputs  = (List) o;
                List<String> results = new ArrayList<String>();

                for (Object input : inputs)
                    results.add(process(input));
                aRow.put(column, results);
            } else if (o instanceof Blob){

                Blob blob = (Blob)o;
                aRow.put(column, readFromBlob(blob));
            }
        }
        return aRow;
    }

    private String process(Object value) {

        if (value == null) return null;
        byte[] bdata = (byte[]) value;
        return new String(bdata);
    }

    private String readFromBlob(Blob blob) {
        try{
            InputStream is = blob.getBinaryStream();
            BufferedReader br = new BufferedReader(new InputStreamReader(is,"UTF-8"));
            String str = "";
            String res = "";
            while((str=br.readLine())!=null){
                res += str;
            }
            return res;
        }catch (Exception e) {
            e.printStackTrace();
            return "";
        }
    }
}

步骤二：
编译源代码。在solr-6.1.0\solr目录下运行ant server

步骤三：

配置db_data_config.xml文件

<dataConfig>
    <dataSource name="jdbcds" type="JdbcDataSource" driver="oracle.jdbc.driver.OracleDriver" url="jdbc:oracle:thin:@//192.168.60.144:1521/NPCBJ" user="user" password="password" />
    <document>
        <entity dataSource="jdbcds" name="proposal" query="select * from BCC_PROPOSAL_INTELLIGENT_VIEW" deltaQuery="select PROPOSAL_ID from BCC_PROPOSAL_INTELLIGENT_VIEW where to_char(modify_time,'yyyy-mm-dd hh24:mi:ss') > '${dataimporter.last_index_time}'"                 
            deltaImportQuery="select * from BCC_PROPOSAL_INTELLIGENT_VIEW where PROPOSAL_ID='${dataimporter.delta.PROPOSAL_ID}'"  
            convertType="true" transformer="BlobTransformer">
                    <field column="PROPOSAL_ID" name="id" />
                    <field column="TITLE" name="title" />   
                    <field column="CONTENT" name="content" blob="true" />
        </entity>
    </document>
</dataConfig>

如此，就可以支持Blob字段了。为了支持她，不容易啊~

参考文章：
Blob values in my table are added to the Solr document as object strings like B@1f23c5
How to index blob field in Apache Solr indexing