hbase的api操作之过滤器

// LESS 小于
//LESS_OR_EQUAL 小于等于
//EQUAL 等于
//NOT_EQUAL 不等于
//GREATER_OR_EQUAL 大于等于
//GREATER 大于
//NO_OP 排除所有

//BinaryComparator 使用Bytes.compareTo()比较
//BinaryPrefixComparator 和BinaryComparator差不多，从前面开始比较
//NullComparator Does not compare against an actual value but whether a given one is null, or not null.
//BitComparator Performs a bitwise comparison, providing a BitwiseOp class with AND, OR, and XOR operators.
// RegexStringComparator 正则表达式
// SubstringComparator 把数据当成字符串，用contains()来判断

//提取rowkey以包含201407的数据
//Filter filter = new RowFilter(CompareFilter.CompareOp.EQUAL,new SubstringComparator("201407"));
// 提取rowkey以01结尾数据
// Filter filter = new RowFilter(CompareFilter.CompareOp.EQUAL,new RegexStringComparator(".*01$"));
//提取rowkey以123开头的数据
//Filter filter = new RowFilter(CompareFilter.CompareOp.EQUAL,new BinaryPrefixComparator("123".getBytes()));

//组合过滤器，且满足至少一条条件

String[] strs = new String[]{"123456", "234567"}

List filters = new ArrayList[Filter]()

for(String str: strs ) {

      RowFilter  filter = new RowFilter(CompareFilter.CompareOp.EQUAL, new BinaryComparator(Bytes.toBytes(cookieid)))

      filters.add(filter)

}

FilterListfilterList = new FilterList(Operator.MUST_PASS_ONE,filters)

/*********************************************以下是各种过滤器的用例**************************************************************/

Comparison Filter：   对比过滤器：
    1.RowFilter
        select * from ns1:t1 where rowkey <= row100
    /**
     * 测试缓存和批处理
     */
    @Test
    public void testBatchAndCaching() throws IOException {

        Configuration conf = HBaseConfiguration.create();
        Connection conn = ConnectionFactory.createConnection(conf);
        TableName tname = TableName.valueOf("ns1:t7");
        Scan scan = new Scan();
        scan.setCaching(2);
        scan.setBatch(4);
        Table t = conn.getTable(tname);
        ResultScanner rs = t.getScanner(scan);
        Iterator<Result> it = rs.iterator();
        while (it.hasNext()) {
            Result r = it.next();
            System.out.println("========================================");
            //得到一行的所有map,key=f1,value=Map<Col,Map<Timestamp,value>>
            NavigableMap<byte[], NavigableMap<byte[], NavigableMap<Long, byte[]>>> map = r.getMap();
            //
            for (Map.Entry<byte[], NavigableMap<byte[], NavigableMap<Long, byte[]>>> entry : map.entrySet()) {
                //得到列族
                String f = Bytes.toString(entry.getKey());
                Map<byte[], NavigableMap<Long, byte[]>> colDataMap = entry.getValue();
                for (Map.Entry<byte[], NavigableMap<Long, byte[]>> ets : colDataMap.entrySet()) {
                    String c = Bytes.toString(ets.getKey());
                    Map<Long, byte[]> tsValueMap = ets.getValue();
                    for (Map.Entry<Long, byte[]> e : tsValueMap.entrySet()) {
                        Long ts = e.getKey();
                        String value = Bytes.toString(e.getValue());
                        System.out.print(f + "/" + c + "/" + ts + "=" + value + ",");
                    }
                }
            }
            System.out.println();
        }
    }

    /**
     * 测试RowFilter过滤器
     */
    @Test
    public void testRowFilter() throws IOException {

        Configuration conf = HBaseConfiguration.create();
        Connection conn = ConnectionFactory.createConnection(conf);
        TableName tname = TableName.valueOf("ns1:t1");
        Scan scan = new Scan();
        RowFilter rowFilter = new RowFilter(CompareFilter.CompareOp.LESS_OR_EQUAL, new BinaryComparator(Bytes.toBytes("row0100")));
        scan.setFilter(rowFilter);
        Table t = conn.getTable(tname);
        ResultScanner rs = t.getScanner(scan);
        Iterator<Result> it = rs.iterator();
        while (it.hasNext()) {
            Result r = it.next();
            System.out.println(Bytes.toString(r.getRow()));
        }
    }

    /**
     * 测试FamilyFilter过滤器
     */
    @Test
    public void testFamilyFilter() throws IOException {

        Configuration conf = HBaseConfiguration.create();
        Connection conn = ConnectionFactory.createConnection(conf);
        TableName tname = TableName.valueOf("ns1:t7");
        Scan scan = new Scan();
        FamilyFilter filter = new FamilyFilter(CompareFilter.CompareOp.LESS, new BinaryComparator(Bytes.toBytes("f2")));
        scan.setFilter(filter);
        Table t = conn.getTable(tname);
        ResultScanner rs = t.getScanner(scan);
        Iterator<Result> it = rs.iterator();
        while (it.hasNext()) {
            Result r = it.next();
            byte[] f1id = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("id"));
            byte[] f2id = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("id"));
            System.out.println(f1id + " : " + f2id);
        }
    }

    /**
     * 测试QualifierFilter(列过滤器)
     */
    @Test
    public void testColFilter() throws IOException {

        Configuration conf = HBaseConfiguration.create();
        Connection conn = ConnectionFactory.createConnection(conf);
        TableName tname = TableName.valueOf("ns1:t7");
        Scan scan = new Scan();
        QualifierFilter colfilter = new QualifierFilter(CompareFilter.CompareOp.EQUAL, new BinaryComparator(Bytes.toBytes("id")));
        scan.setFilter(colfilter);
        Table t = conn.getTable(tname);
        ResultScanner rs = t.getScanner(scan);
        Iterator<Result> it = rs.iterator();
        while (it.hasNext()) {
            Result r = it.next();
            byte[] f1id = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("id"));
            byte[] f2id = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("id"));
            byte[] f2name = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("name"));
            System.out.println(f1id + " : " + f2id + " : " + f2name);
        }
    }


    /**
     * 测试ValueFilter(值过滤器)
     * 过滤value的值，含有指定的字符子串
     */
    @Test
    public void testValueFilter() throws IOException {

        Configuration conf = HBaseConfiguration.create();
        Connection conn = ConnectionFactory.createConnection(conf);
        TableName tname = TableName.valueOf("ns1:t7");
        Scan scan = new Scan();
        ValueFilter filter = new ValueFilter(CompareFilter.CompareOp.EQUAL, new SubstringComparator("to"));
        scan.setFilter(filter);
        Table t = conn.getTable(tname);
        ResultScanner rs = t.getScanner(scan);
        Iterator<Result> it = rs.iterator();
        while (it.hasNext()) {
            Result r = it.next();
            byte[] f1id = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("id"));
            byte[] f2id = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("id"));
            byte[] f1name = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("name"));
            byte[] f2name = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("name"));
            System.out.println(f1id + " : " + f2id + " : " + Bytes.toString(f1name) + " : " + Bytes.toString(f2name));
        }
    }

    /**
     * DependentColumnFilter 依赖列过滤器     自定义一种过滤基数
     */
    @Test
    public void testDepFilter() throws IOException {
        Configuration conf = HBaseConfiguration.create();
        Connection conn = ConnectionFactory.createConnection(conf);
        TableName tname = TableName.valueOf("ns1:t7");
        Scan scan = new Scan();
        DependentColumnFilter filter = new DependentColumnFilter(Bytes.toBytes("f2"),
                Bytes.toBytes("addr"),
                true,
                CompareFilter.CompareOp.NOT_EQUAL,
                new BinaryComparator(Bytes.toBytes("beijing"))
                );
        scan.setFilter(filter);
        Table t = conn.getTable(tname);
        ResultScanner rs = t.getScanner(scan);
        Iterator<Result> it = rs.iterator();
        while (it.hasNext()) {
            Result r = it.next();
            byte[] f1id = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("id"));
            byte[] f2id = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("id"));
            byte[] f1name = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("name"));
            byte[] f2name = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("name"));
            System.out.println(f1id + " : " + f2id + " : " + Bytes.toString(f1name) + " : " + Bytes.toString(f2name));
        }
    }

    /**
     * 单列值value过滤，  和值过滤器对比看
     * 如果value不满足，整行滤掉
     */
    @Test
    public void testSingleColumValueFilter() throws IOException {

        Configuration conf = HBaseConfiguration.create();
        Connection conn = ConnectionFactory.createConnection(conf);
        TableName tname = TableName.valueOf("ns1:t7");
        Scan scan = new Scan();
        SingleColumnValueFilter filter = new SingleColumnValueFilter(Bytes.toBytes("f2",
                Bytes.toBytes("name"),
                CompareFilter.CompareOp.NOT_EQUAL),
                new BinaryComparator(Bytes.toBytes("tom2.1")));

        //ValueFilter filter = new ValueFilter(CompareFilter.CompareOp.EQUAL, new SubstringComparator("to"));
        scan.setFilter(filter);
        Table t = conn.getTable(tname);
        ResultScanner rs = t.getScanner(scan);
        Iterator<Result> it = rs.iterator();
        while (it.hasNext()) {
            Result r = it.next();
            byte[] f1id = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("id"));
            byte[] f2id = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("id"));
            byte[] f1name = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("name"));
            byte[] f2name = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("name"));
            System.out.println(f1id + " : " + f2id + " : " + Bytes.toString(f1name) + " : " + Bytes.toString(f2name));
        }
    }
复杂查询  like
-----------------
    如 sql：  select * from t7 where ((age <= 13) and (name like '%t')  or  (age > 13) and (name like 't%'))
FilterList
----------------
    @Test
    public void testComboFilter() throws IOException {

        Configuration conf = HBaseConfiguration.create();
        Connection conn = ConnectionFactory.createConnection(conf);
        TableName tname = TableName.valueOf("ns1:t7");
        Scan scan = new Scan();

        //where ... f2:age <= 13
        SingleColumnValueFilter ftl = new SingleColumnValueFilter(
                Bytes.toBytes("f2"),
                Bytes.toBytes("age"),
                CompareFilter.CompareOp.LESS_OR_EQUAL,
                new BinaryComparator(Bytes.toBytes("13"))
        );

        //where ... f2:name like %t
        SingleColumnValueFilter ftr = new SingleColumnValueFilter(
                Bytes.toBytes("f2"),
                Bytes.toBytes("name"),
                CompareFilter.CompareOp.EQUAL,
                new RegexStringComparator("^t")
        );
        //ft
        FilterList ft = new FilterList(FilterList.Operator.MUST_PASS_ALL);
        ft.addFilter(ftl);
        ft.addFilter(ftr);

        //where ... f2:age > 13
        SingleColumnValueFilter fbl = new SingleColumnValueFilter(
                Bytes.toBytes("f2"),
                Bytes.toBytes("age"),
                CompareFilter.CompareOp.GREATER,
                new BinaryComparator(Bytes.toBytes("13"))
        );

        //where ... f2:name like %t
        SingleColumnValueFilter fbr = new SingleColumnValueFilter(
                Bytes.toBytes("f2"),
                Bytes.toBytes("name"),
                CompareFilter.CompareOp.EQUAL,
                new RegexStringComparator("t$")
        );
        //ft
        FilterList fb = new FilterList(FilterList.Operator.MUST_PASS_ALL);
        fb.addFilter(fbl);
        fb.addFilter(fbr);


        FilterList fall = new FilterList(FilterList.Operator.MUST_PASS_ONE);
        fall.addFilter(ft);
        fall.addFilter(fb);

        scan.setFilter(fall);
        Table t = conn.getTable(tname);
        ResultScanner rs = t.getScanner(scan);
        Iterator<Result> it = rs.iterator();
        while (it.hasNext()) {
            Result r = it.next();
            byte[] f1id = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("id"));
            byte[] f2id = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("id"));
            byte[] f1name = r.getValue(Bytes.toBytes("f1"), Bytes.toBytes("name"));
            byte[] f2name = r.getValue(Bytes.toBytes("f2"), Bytes.toBytes("name"));
            System.out.println(f1id + " : " + f2id + " : " + Bytes.toString(f1name) + " : " + Bytes.toString(f2name));
        }
    }

分页过滤器：

列分页过滤:
 select "fl:name"
keyOnly过滤器 ，只提取key 。丢弃value；