Hive自定义函数UDF(统计字符串个数)

需求

统计一个字符串在另一个字符串中出现的次数，比如："a or b or c"中出现了多少个"or"

分析

Hive内置函数中并没有可以直接使用的函数，嵌套起来比较麻烦，所以需要自定义实现

代码

package udf;

import org.apache.hadoop.hive.ql.exec.UDFArgumentException;
import org.apache.hadoop.hive.ql.exec.UDFArgumentLengthException;
import org.apache.hadoop.hive.ql.metadata.HiveException;
import org.apache.hadoop.hive.ql.udf.generic.GenericUDF;
import org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector;
import org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory;
import org.apache.hadoop.hive.serde2.objectinspector.primitive.StringObjectInspector;

/**
 * @ClassName: UDFCntStr
 * @Description: 
 * @Author: xuezhouyi
 * @Version: V1.0
 **/
public class UDFCntStr extends GenericUDF {
	StringObjectInspector line;
	StringObjectInspector word;

	@Override
	public ObjectInspector initialize(ObjectInspector[] arguments) throws UDFArgumentException {
		if (arguments.length != 2) {
			throw new UDFArgumentLengthException("arrayContainsExample only takes 2 arguments: T, T");
		}
		
		/* 检查是否接收到正确的参数类型 */
		ObjectInspector a = arguments[0];
		ObjectInspector b = arguments[1];
		if (!(a instanceof StringObjectInspector) || !(b instanceof StringObjectInspector)) {
			throw new UDFArgumentException("first argument must be a string, second argument must be a string");
		}
		this.line = (StringObjectInspector) a;
		this.word = (StringObjectInspector) b;

		/* 返回类型是int，所以我们提供了正确的object inspector */
		return PrimitiveObjectInspectorFactory.javaIntObjectInspector;
	}

	@Override
	public Object evaluate(DeferredObject[] arguments) throws HiveException {
		/* 利用object inspectors从传递的对象中得到line与word */
		String line = this.line.getPrimitiveJavaObject(arguments[0].get());
		String word = this.word.getPrimitiveJavaObject(arguments[1].get());

		/* 递归调用统计个数 */
		return fun(line, word, 0, 0);
	}

	@Override
	public String getDisplayString(String[] children) {
		return "line count word example";
	}

	/* 递归函数 */
	private static int fun(String s1, String s2, int start, int count) {
		/* 从开始位置截取 */
		s1 = s1.substring(start);

		/* 判断是否存在s2不存在则直接返回count */
		if ((start = s1.indexOf(s2)) == -1)
			return count;

		/* 存在则加一，然后开始值后移s2的长度作为下一次的开始值 */
		count++;
		start += s2.length();

		/* 递归 */
		return fun(s1, s2, start, count);
	}
}

Maven打包上传添加jar包

add jar /home/hadoop/tmp/hive-1.0-SNAPSHOT.jar;

创建临时函数

create temporary function UDFCntStr as 'udf.UDFCntStr';

最终效果