Python+SQL：将Dataframe上传到数据库_python dataframe上传到数据-优快云博客

* 上传数据前需确认清楚自己是否有权限访问host，以及确认清楚具体的上传路径（hdfs_path）

import pandas as pd
from impala.dbapi import connect
from hdfs.client import Client
from hdfs import InsecureClient
import pyhdfs

def getDatabaseConnection():
    """
    获取数据库连接和游标
    """
    return connect(host=XX, port=XX, auth_mechanism=XX, user=XX, password=XX, database=XX)

def executeSqlString(sql_string):
    """
    执行SQL语句
    """
    with getDatabaseConnection().cursor() as cur:
        cur.execute(sql_string)
    return 'SQL已执行'

def uploadDataframe2Database(table_name, create_table_sql, df_to_upload):
    # pick active node
    fs = pyhdfs.HdfsClient(hosts=[http1,http2,http3])
    active_nodes=fs.get_active_namenode()

    # Drop existing table
    drop_table_sql = f'DROP TABLE IF EXISTS {table_name}'
    executeSqlString(drop_table_sql)

    # Create new table
    executeSqlString(create_table_sql)

    # Upload DataFrame to HDFS
    hdfs_path = XX
    client = InsecureClient(active_nodes)
    with client.write(hdfs_path, overwrite=True, encoding='utf-8') as writer:
        df_to_upload.to_csv(writer, index=False, header=False, sep='\001')
    return f'{table_name} 数据已上传！'

df_test = pd.DataFrame({'class':['A','B','C'],'rank':[3,2,3]})
table_name = table_name
create_table_sql = f'''
CREATE TABLE table_name(
    class STRING,
    rank INT
)
row format delimited fields terminated by '\001' lines terminated by '\n' STORED AS textfile
'''

print(uploadDataframe2Database(table_name,create_table_sql,df_test))

数据上传请看：Python+SQL：在python中调用sql文件并运行转化为Datafarme-优快云博客