HBase中使用过滤器筛选数据

骑猪看日落 2023-02-12 05:55 45阅读 0赞

## 一、过滤器能干什么 ##

*  HBase为筛选数据提供了一组过滤器，通过过滤器可以在HBase中的数据的多个维度(行，列，数据版本)上进行对数据的筛选操作。
 *  通常来说，通过行键、列来筛选数据的应用场景较多。

--------------------

## 二、常见的过滤器 ##

1.  基于行的过滤器
    
     *  PrefixFilter: 行的前缀匹配
     *  PageFilter: 基于行的分页
2.  基于列的过滤器
    
     *  ColumnPrefixFilter: 列前缀匹配
     *  FirstKeyOnlyFilter: 只返回每一行的第一列
3.  基于单元值的过滤器
    
     *  KeyOnlyFilter: 返回的数据不包括单元值，只包含行键与列
     *  TimestampsFilter: 根据数据的时间戳版本进行过滤
4.  基于列和单元值的过滤器
    
     *  SingleColumnValueFilter: 对该列的单元值进行比较过滤
     *  SingleColumnValueExcludeFilter: 对该列的单元值进行比较过滤
5.  比较过滤器
    
     *  比较过滤器通常需要一个比较运算符以及一个比较器来实现过滤
     *  RowFilter、 FamilyFilter、 QualifierFilter、 ValueFilter
6.  常见过滤器总结

<table> 
 <thead> 
  <tr> 
   <th>过滤器(Filter)</th> 
   <th>功能</th> 
  </tr> 
 </thead> 
 <tbody> 
  <tr> 
   <td>RowFilter</td> 
   <td>筛选出匹配的所有的行</td> 
  </tr> 
  <tr> 
   <td>PrefixFilter</td> 
   <td>筛选出具有特定前缀的行键的数据</td> 
  </tr> 
  <tr> 
   <td>KeyOnlyFilter</td> 
   <td>只返回每行的行键，值全部为空</td> 
  </tr> 
  <tr> 
   <td>ColumnPrefixFilter</td> 
   <td>按照列名的前缀来筛选单元格</td> 
  </tr> 
  <tr> 
   <td>ValueFilter</td> 
   <td>按照具体的值来筛选单元格的过滤器</td> 
  </tr> 
  <tr> 
   <td>TimestampsFilter</td> 
   <td>根据数据的时间戳版本进行过滤</td> 
  </tr> 
  <tr> 
   <td>FilterList</td> 
   <td>用于综合使用多个过滤器</td> 
  </tr> 
 </tbody> 
</table>

--------------------

## 三、开发演示 ##

/** * @title HBaseFilterTest * @date 2019/12/9 15:01 * @description 尝试使用过滤器 */
    public class HBaseFilterTest { 
    
        @Test
        public void createTable(){ 
            HBaseUtil.createTable("FileTable", new String[]{ "fileInfo", "saveInfo"});
        }
    
        @Test
        public void addFileDetails(){ 
            HBaseUtil.putRow("FileTable", "rowkey1", "fileInfo", "name", "file1.txt");
            HBaseUtil.putRow("FileTable", "rowkey1", "fileInfo", "type", "txt");
            HBaseUtil.putRow("FileTable", "rowkey1", "fileInfo", "size", "1024");
            HBaseUtil.putRow("FileTable", "rowkey1", "saveInfo", "creator", "suiwo1");
            HBaseUtil.putRow("FileTable", "rowkey2", "fileInfo", "name", "file2.jpg");
            HBaseUtil.putRow("FileTable", "rowkey2", "fileInfo", "type", "jpg");
            HBaseUtil.putRow("FileTable", "rowkey2", "fileInfo", "size", "2048");
            HBaseUtil.putRow("FileTable", "rowkey2", "saveInfo", "creator", "suiwo3");
            HBaseUtil.putRow("FileTable", "rowkey3", "fileInfo", "name", "file3.jpg");
            HBaseUtil.putRow("FileTable", "rowkey3", "fileInfo", "type", "jpg");
            HBaseUtil.putRow("FileTable", "rowkey3", "fileInfo", "size", "2048");
            HBaseUtil.putRow("FileTable", "rowkey3", "saveInfo", "creator", "suiwo3");
        }
    
        /** * rowkey = rowkey1 * fileName = file1.txt */
        @Test
        public void rowFilterTest(){ 
            Filter filter = new RowFilter(CompareFilter.CompareOp.EQUAL, new BinaryComparator(Bytes.toBytes("rowkey1")));
    
            // MUST_PASS_ALL指必须通过所有的Filter
            FilterList filterList = new FilterList(FilterList.Operator.MUST_PASS_ALL, Arrays.asList(filter));
    
            ResultScanner scanner = HBaseUtil.getScanner("FileTable","rowkey1","rowkey3", filterList);
    
            if(scanner != null){ 
                scanner.forEach(result -> { 
                    System.out.println("rowkey = " + Bytes.toString(result.getRow()));
                    System.out.println("fileName = " + Bytes.toString(result.getValue(Bytes.toBytes("fileInfo"), Bytes.toBytes("name"))));
                });
                scanner.close();
            }
        }
    
        /** * rowkey = rowkey2 * fileName = file2.jpg */
        @Test
        public void prefixFilterTest(){ 
            Filter filter = new PrefixFilter(Bytes.toBytes("rowkey2"));
            FilterList filterList = new FilterList(FilterList.Operator.MUST_PASS_ALL, Arrays.asList(filter));
            ResultScanner scanner = HBaseUtil.getScanner("FileTable","rowkey1","rowkey3", filterList);
    
            if(scanner != null){ 
                scanner.forEach(result -> { 
                    System.out.println("rowkey = " + Bytes.toString(result.getRow()));
                    System.out.println("fileName = " + Bytes.toString(result.getValue(Bytes.toBytes("fileInfo"), Bytes.toBytes("name"))));
                });
                scanner.close();
            }
        }
    
        /** * rowkey = rowkey1 * fileName = ��� * rowkey = rowkey2 * fileName = ��� */
        @Test
        public void keyOnlyFilterTest(){ 
            Filter filter = new KeyOnlyFilter(true);
            FilterList filterList = new FilterList(FilterList.Operator.MUST_PASS_ALL, Arrays.asList(filter));
            ResultScanner scanner = HBaseUtil.getScanner("FileTable","rowkey1","rowkey3", filterList);
    
            if(scanner != null){ 
                scanner.forEach(result -> { 
                    System.out.println("rowkey = " + Bytes.toString(result.getRow()));
                    System.out.println("fileName = " + Bytes.toString(result.getValue(Bytes.toBytes("fileInfo"), Bytes.toBytes("name"))));
                });
                scanner.close();
            }
        }
    
        /** * rowkey = rowkey1 * fileName = file1.txt * fileType = null * rowkey = rowkey2 * fileName = file2.jpg * fileType = null */
        @Test
        public void columnPrefixFilterTest(){ 
            Filter filter = new ColumnPrefixFilter(Bytes.toBytes("nam"));// 前缀为nam
            FilterList filterList = new FilterList(FilterList.Operator.MUST_PASS_ALL, Arrays.asList(filter));
            ResultScanner scanner = HBaseUtil.getScanner("FileTable","rowkey1","rowkey3", filterList);
    
            if(scanner != null){ 
                scanner.forEach(result -> { 
                    System.out.println("rowkey = " + Bytes.toString(result.getRow()));
                    System.out.println("fileName = " + Bytes.toString(result.getValue(Bytes.toBytes("fileInfo"), Bytes.toBytes("name"))));
                    System.out.println("fileType = " + Bytes.toString(result.getValue(Bytes.toBytes("fileInfo"), Bytes.toBytes("type"))));
                });
                scanner.close();
            }
        }
    }