![](/img/trans.png)
[英]Python async redis gives error AttributeError: __aexit__
[英]python asynchronous data pull error: __aexit__/__enter__
我正在嘗試使用 Python 的異步功能編寫代碼。 我有一個數據庫連接 class ,其中我有用於(斷開)連接數據庫以及獲取數據的代碼。 現在我想使用基於一個標識符的 fetch data 方法異步獲取數據。 代碼如下圖:
import pyexasol
import pandas as pd
import logging
from typing import Iterable
import asyncio
import tqdm
class Exa(object):
def __init__(self, dsn: str = '1.2.3.4',
user: str = os.environ['UID'],
password: str = os.environ['PWD']):
self.__dsn = dsn
self.__user = user
self.__password = password
self.conn = None
def __connect(self):
if self.conn is None:
try:
self.conn = pyexasol.connect(dsn=self.__dsn, user=self.__user,
password=self.__password, encryption=True)
except Exception as e:
logging.error(f"Error in connecting with Exasol. Error is: {e}")
def __disconnect(self):
if self.conn is not None:
try:
self.conn.close()
except Exception as e:
logging.error(f"Exception in disconnecting DB. Error is {e}")
self.conn = None
def fetch(self, query: str, leave_connection_open: bool = False) -> pd.DataFrame:
# connect and execute the query
self.__connect()
try:
res = self.conn.export_to_pandas(query)
res.columns = res.columns.str.lower()
except Exception as e:
self.__disconnect()
return pd.DataFrame()
if not leave_connection_open:
self.__disconnect()
return res
def fetch_batch(self, pattern: str, replacement: Iterable,
query: str, batchsize: int = 5000) -> pd.DataFrame:
res = asyncio.run(self._fetch_batch(pattern=pattern, replacement=replacement,
query=query, batchsize=batchsize))
return res
async def _fetch_batch(self, pattern: str, replacement: Iterable,
query: str, batchsize: int = 5000) -> pd.DataFrame:
replacement = list(replacement)
# breaking into batches
if any(isinstance(i, str) for i in replacement):
batches = ["'" + "','".join(replacement[i:i + batchsize]) + "'"
for i in range(0, len(replacement), batchsize)]
else:
batches = [",".join(replacement[i:i + batchsize])
for i in range(0, len(replacement), batchsize)]
# connecting and executing query in batches
nbatches = len(batches)
self.__connect()
try:
tasks = [self.__run_batch_query(query=query.replace(pattern, batches[i]),
i=i, nbatches=nbatches) for i in range(nbatches)]
# progress bar
res = [await f for f in tqdm.tqdm(asyncio.as_completed(tasks), total=len(tasks))]
except Exception as e:
logging.error("Could not fetch batches of data. Error is: %s", e)
'''finally:
self.__disconnect()'''
# dataframe concatenation
res = pd.concat(res)
res.columns = res.columns.str.lower()
return res
async def __run_batch_query(self, query: str,
i: int, nbatches: int) -> pd.DataFrame:
logging.info("Fetching %d/%d", i + 1, nbatches)
async with self.fetch(query=query, leave_connection_open=True) as resp:
raw = await resp
return raw
我正在運行此代碼:
from foo import Exa
db = Exa()
ids = db.fetch('select id from application limit 100')
ids1 = db.fetch_batch(pattern='IDS',
replacement=ids['id'],
query='select id from application where id in (IDS)',
batchsize=25)
但后來我得到如下錯誤:
ERROR:root:Could not fetch batches of data. Error is: __aexit__
Traceback (most recent call last):
File "/home/priya/pydbutils/gitignored/foo2.py", line 85, in __run_batch_query
async with self.fetch(query=query, leave_connection_open=True) as resp:
AttributeError: __aexit__
此外,如果我在沒有async
的情況下將__run_batch_query()
方法調用更改為self.fetch()
方法,則錯誤將更改為:
ERROR:root:Could not fetch batches of data. Error is: __enter__
Traceback (most recent call last):
File "/home/priya/pydbutils/gitignored/foo2.py", line 85, in __run_batch_query
with self.fetch(query=query, leave_connection_open=True) as resp:
AttributeError: __enter__
如果這里有錯誤,請幫忙指出錯誤?
pyexasol 創造者在這里。
請注意,asyncio 不會為與 Exasol 相關的場景提供任何好處。 Ayncio 在單個 CPU 上運行並使用單個網絡連接,這會阻止任何有意義的擴展。
從 Exasol 服務器加載數據的最有效方法是:
export_to_pandas()
或export_to_callback()
用於單個 Python 進程;export_parallel()
+ http_transport()
用於多個 Python 進程; 請查看HTTP Transport (parallel)
手冊頁以獲取解釋和示例。 這種方法可以線性擴展,您甚至可以在多個服務器上運行計算任務。
對於簡單的場景,如果您通過慢速(例如 WiFi)網絡傳輸大量數據,您可以考慮使用compression=True
連接選項。
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.