我尝试使用sqlalchemy
通过pandas附加我的vertica(SQL类型)表import pandas as pd
import sqlalchemy as sa
创建引擎到vertica:
def get_engine(base):
engine = sa.create_engine("{sys}+{dri}://{user}:" + \
"{password}@{host}:{port}/{database}".format(**login[base]))
return engine
engine = get_engine('vertica')
为了清楚起见,一个简单的查询:
table = '***'
sql =\
'''
select *
from public.{table}
'''.format(table=table)
connection = engine.connect()
data = pd.read_sql(sql, connection)
connection.close()
数据不为空:
print(len(data))
569955
并尝试写入同一张表:
fields = list(data.columns)
connection = engine.connect()
data.to_sql(table, connection, schema='public', index=False, if_exists='append', chunksize=30000,
dtype={fields[0]:sa.types.Integer,
fields[1]:sa.types.VARCHAR,
fields[2]:sa.types.Integer,
fields[3]:sa.types.Integer,
fields[4]:sa.types.Integer,
fields[5]:sa.types.VARCHAR,
fields[6]:sa.types.VARCHAR,
fields[7]:sa.types.VARCHAR,
fields[8]:sa.types.VARCHAR,
fields[9]:sa.types.VARCHAR,
fields[10]:sa.types.VARCHAR,
fields[11]:sa.types.VARCHAR,
fields[12]:sa.types.DateTime
})
connection.close()
并犯了这个错误:
...
\Anaconda3\lib\site-packages\sqlalchemy\engine\default.py in do_executemany(self, cursor, statement, parameters, context)
465
466 def do_executemany(self, cursor, statement, parameters, context=None):
--> 467 cursor.executemany(statement, parameters)
468
469 def do_execute(self, cursor, statement, parameters, context=None):
\Anaconda3\lib\site-packages\vertica_python\vertica\cursor.py in executemany(self, operation, seq_of_parameters)
153 else:
154 raise NotImplementedError(
--> 155 "executemany is implemented for simple INSERT statements only")
156
157 def fetchone(self):
NotImplementedError: executemany is implemented for simple INSERT statements only
答案 0 :(得分:1)
当我尝试使用sqlalchemy将数据写入vertica时,出现了相同的错误。就我而言,问题是列名。似乎无法写出包含特殊字符的列名。我可以通过从pandas的列名称中删除所有的'_','%'和空格字符来解决错误,然后使用df.to_sql()将其写入vertica。