python使用jdbc连接hive

At last,I figure it out finally. But I think is not a great idea to use this component to connect hive, it's depend lots of jar package, and you must install jar package with python-java debug REPL(mostly you are guessing),it is bad.

refer : https://github.com/ghostyusheng/python-hive-jdbc

# -*- coding: utf8 -*-

import jaydebeapi
import os


class Jdbc:
    base_sql = [
        "set hive.mapred.mode=nonstrict",
        "set hive.strict.checks.cartesian.product=false",
        "set hive.execution.engine=tez"
    ]

    def query(self, sql, db = 'tap_ods'):  
        url = 'jdbc:hive2://xx.xx:10000/' + db
        dirver = 'org.apache.hive.jdbc.HiveDriver'
        DIR = os.getcwd() + '/lib/'
        jarFile = [
            DIR + 'hive-jdbc-3.1.1.jar',
            DIR + 'commons-logging-1.2.jar',
            DIR + 'hive-service-3.1.1.jar',
            DIR + 'hive-service-rpc-3.1.1.jar',
            DIR + 'libthrift-0.12.0.jar',
            DIR + 'httpclient-4.5.9.jar',
            DIR + 'httpcore-4.4.11.jar',
            DIR + 'slf4j-api-1.7.26.jar',
            DIR + 'curator-framework-4.2.0.jar',
            DIR + 'curator-recipes-4.2.0.jar',
            DIR + 'curator-client-4.2.0.jar',
            DIR + 'commons-lang-2.6.jar',
            DIR + 'hadoop-common-3.2.0.jar',
            DIR + 'httpcore-4.4.11.jar',
            DIR + 'hive-common-3.1.1.jar',
            DIR + 'hive-serde-3.1.1.jar',
            DIR + 'guava-28.0-jre.jar'
        ]
        conn = jaydebeapi.connect(dirver, url, ['hadoop', ''], jarFile)
        curs = conn.cursor()
        for _sql in self.base_sql:
            curs.execute(_sql)
        curs.execute(sql)
        result = curs.fetchall()
        curs.close()
        conn.close()
        return result

Notice:

1.JPype1 version use 0.6.3 -> pip3 install JPype1==0.6.3

2.pip3 install Jaydebeapi==1.1.1

3.makesure hive2 OR hive?

use code , Download jar package at https://mvnrepository.com/