java - 失败:SemanticException [错误10014]:行1:21错误的参数 'stock_price_high':

Question

我是Hive的新手，我正在尝试在Hive命令行中创建和使用UDF。

我创建了一个Java代码来根据样本NYSE数据集计算股票的协方差。以下是Java中的代码:

package udf;

import org.apache.hadoop.hive.ql.exec.UDF;

public class CoVariance extends UDF {
    public Double covariance (Double stockpricex, Double stockpricey, Double avgstockpricex, Double avgstockpricey, int tuplecount) {

        if (stockpricex == null|| stockpricey == null || avgstockpricex == null || avgstockpricey == null || tuplecount == 0 ) {
        return null;
        } //check for invalid parameters

        Double covar = ((stockpricex-avgstockpricex)*(stockpricey-avgstockpricey)/(tuplecount - 1));

        return covar.doubleValue();     
    } // return the final co-variance of the stocks

} // end of class

将jar文件添加到配置单元 shell 后，我创建了一个名为“cv”的临时函数并传递了参数。

select stock_symbol, cv(stock_price_high, stock_price_low, avg(stock_price_high), avg(stock_price_low), count(stock_price_high)) from nyse group by stock_symbol, stock_price_high, stock_price_low;

我得到以下错误:

FAILED: SemanticException [Error 10014]: Line 1:21 Wrong arguments 'stock_price_high': No matching method for class udf.CoVariance with (double, double, double, double, bigint). Possible choices:

这是怎么了？您的帮助将不胜感激。

Answer 1

不清楚如何注册UDF，但是JavaDoc表示

Requirements for all classes extending this UDF are:
- Implement one or more methods named evaluate which will be called by Hive

您的方法 covariance应该重命名为 evaluate
根据Hive错误， int tuplecount应该是 long或 BigInteger