python连接Greenplum数据库

pip install psycopg2

psycopg2.connect(host="localhost", port="5432", dbname="testdb", user="gpadmin", password="123456")

import psycopg2
import psycopg2.extras
import time

‘‘‘
    连接数据库
    returns:db
‘‘‘
def gp_connect():
    try:
        db = psycopg2.connect(dbname="testdb",
                              user="gpadmin",
                              password="gpadmin",
                              host="10.1.208.42",
                              port="5432")
        # connect()也可以使用一个大的字符串参数,
        # 比如”host=localhost port=5432 user=postgres password=postgres dbname=test”
        return db
    except psycopg2.DatabaseError as e:
        print("could not connect to Greenplum server",e)


if __name__ == ‘__main__‘:
    conn = gp_connect()
    print(conn)
    cur = conn.cursor(cursor_factory=psycopg2.extras.RealDictCursor)
    # 这里创建的是一个字典Cursor, 这样返回的数据, 都是字典的形式, 方便使用
    ret = cur.execute("CREATE TABLE public.gp_test (id serial PRIMARY KEY, num integer, data varchar);")
    conn.commit()
    # 提交到数据库中
    print(ret)
    ret = cur.execute("INSERT INTO public.gp_test (num, data) VALUES (%s, %s);",(300, "abc‘def"))

    conn.commit()
    # 提交到数据库中
    print(cur.rowcount)  # 1
    # 返回数据库中的行的总数已修改，插入或删除最后 execute*().

    ret_sql = cur.mogrify("select * from pg_tables where tablename = %s;", (‘gp_test‘,))
    # 返回生成的sql脚本, 用以查看生成的sql是否正确.
    # sql脚本必须以;结尾, 不可以省略.其次, 不管sql中有几个参数, 都需要用 % s代替, 只有 % s, 不管值是字符还是数字, 一律 % s.
    # 最后, 第二个参数中, 一定要传入元组, 哪怕只有一个元素, 像我刚才的例子一样, (‘gp_test‘)这样是不行的.
    print(ret_sql.decode(‘utf-8‘))  # select * from pg_tables where tablename = E‘gp_test‘;

    cur.execute("select * from gp_test where num = %s;", (300,))
    pg_obj = cur.fetchone()
    print(pg_obj) # {‘id‘: 1, ‘num‘: 300, ‘data‘: "abc‘def"}

    conn.close() # 关闭连接

    conn = gp_connect()
    print(conn)
    cur = conn.cursor(cursor_factory=psycopg2.extras.RealDictCursor)
    # # 这里创建的是一个字典Cursor, 这样返回的数据, 都是字典的形式, 方便使用
    # ret = cur.execute("CREATE TABLE public.gp_test (id serial PRIMARY KEY, num integer, data varchar);")
    # conn.commit()
    # # 提交到数据库中
    # print(ret)
    gp_list = []
    for i in range(200):
        gp_list.append((i,‘abc%s‘%i))
    # print(gp_list)
    # 批量提交数据
    ret = cur.executemany("INSERT INTO public.gp_test (num, data) VALUES (%s, %s);", gp_list)
    conn.commit()
    # 提交到数据库中
    print(cur.query)  # 查看上一条执行的脚本
    print(cur.rowcount)  # 200
    # 返回数据库中的行的总数已修改，插入或删除最后 execute*().
    cur.execute("select  count(*) num from gp_test")
    pg_obj = cur.fetchone()
    print(pg_obj)  # {‘num‘: 200}

    conn.close()  # 关闭连接

import psycopg2
import psycopg2.extras
import psycopg2.pool
from datetime import datetime

‘‘‘
    连接数据库
    使用数据库连接池
    returns:db
‘‘‘
def gp_connect():
    try:
        simple_conn_pool = psycopg2.pool.SimpleConnectionPool(minconn=1, maxconn=5,dbname="testdb",
                              user="gpadmin",
                              password="gpadmin",
                              host="10.1.208.42",
                              port="5432")
        # connect()也可以使用一个大的字符串参数,
        # 比如”host=localhost port=5432 user=postgres password=postgres dbname=test”
        # 从数据库连接池获取连接
        conn = simple_conn_pool.getconn()
        return conn
    except psycopg2.DatabaseError as e:
        print("could not connect to Greenplum server",e)


if __name__ == ‘__main__‘:
    conn = gp_connect()
    print(conn)
    cur = conn.cursor()
    # 批量查询大小
    batch_size = 1000
    gp_list = []
    for i in range(2000, 100000):
        gp_list.append((i,‘abc%s‘%i))
    # print(gp_list)

    # 开始时间
    start_time = datetime.now()
    # 批量提交数据execute_values性能大于executemany
    psycopg2.extras.execute_values(cur, "INSERT INTO public.gp_test (num, data) VALUES %s", gp_list)
    conn.commit()
    # 提交到数据库中
    cur.execute("select  *  from gp_test order by id")
    count = 0

    while True:
        count = count + 1
        # 每次获取时会从上次游标的位置开始移动size个位置，返回size条数据
        data = cur.fetchmany(batch_size)
        # 数据为空的时候中断循环
        if not data:
            break
        else:
            print(data[-1])  # 得到最后一条(通过元祖方式返回)
        print(‘获取%s到%s数据成功‘ % ((count - 1) * batch_size, count * batch_size))
    print(‘insert到fetchmany获取全量数据所用时间:‘, (datetime.now() - start_time).seconds) # 16s
conn.close()  # 关闭连接