HBase学习笔记-统计表行数

HBase统计表行数的方式如下：

一、HBase Shell

hbase> count 'tableName', INTERVAL => 1000, CACHE => 1000 

INTERVAL为统计的行数间隔，默认为1000，CACHE为统计的数据缓存

二、HBase自带MapReduce表行数统计RowCounter

$HBASE_HOME/bin/hbase org.apache.hadoop.hbase.mapreduce.RowCounter  'tableName'

$HBASE_HOME/bin/hbase org.apache.hadoop.hbase.mapreduce.RowCounter  -Dmapreduce.job.queuename=queue1 'tableName'

三、使用Scan与Filter的方式对表行数进行统计

public static long rowCount(String tableName) {
	long rowCount = 0;
	try {
		HTable table = new HTable(configuration, tableName);
		Scan scan = new Scan();
		scan.setFilter(new FirstKeyOnlyFilter());
		ResultScanner resultScanner = table.getScanner(scan);
		for (Result result : resultScanner) {
			rowCount += result.size();
		}
	} catch (IOException e) {
		logger.info(e.getMessage(), e);
	}
	return rowCount;
}

四、使用Coprocessor新特性来对表行数进行统计

public static void addTableCoprocessor(String tableName, String coprocessorClassName) {
	try {
		admin.disableTable(tableName);
		HTableDescriptor htd = admin.getTableDescriptor(Bytes.toBytes(tableName));
		htd.addCoprocessor(coprocessorClassName);
		admin.modifyTable(Bytes.toBytes(tableName), htd);
		admin.enableTable(tableName);
	} catch (IOException e) {
		logger.info(e.getMessage(), e);
	}
}
	
public static long rowCount(String tableName, String family) {
	AggregationClient ac = new AggregationClient(configuration);  
	Scan scan = new Scan();
	scan.addFamily(Bytes.toBytes(family));
	long rowCount = 0;
	try {
		rowCount = ac.rowCount(Bytes.toBytes(tableName), new LongColumnInterpreter(), scan);
	} catch (Throwable e) {
		logger.info(e.getMessage(), e);
	}  
	return rowCount;
}

@Test
public void testTableRowCount() {
	String coprocessorClassName = "org.apache.hadoop.hbase.coprocessor.AggregateImplementation";
	HBaseUtils.addTableCoprocessor("user", coprocessorClassName);
	long rowCount = HBaseUtils.rowCount("user", "basic");
	System.out.println("rowCount: " + rowCount);
}