以下Python代码段说明了此问题:
print("starting")
# I am trying to configure a query/command timeout of one second.
# sqlalchemy docs suggest execution_options but the documented list of options doesn't include a timeout:
# http://docs.sqlalchemy.org/en/latest/core/connections.html#sqlalchemy.engine.Connection.execution_options
# Below, I am guessing at several likely timeout parameter names:
db_engine = create_engine("postgresql://user:pass@server/catalog",
execution_options={"timeout": 1.0,
"statement_timeout": 1.0,
"query_timeout": 1.0,
"execution_timeout": 1.0})
with db_engine.connect() as db_connection:
print("got db_connection")
# Artificially force a two second query time with pg_sleep.
# This is designed to guarantee timeout conditions and trigger an exception.
result_proxy = db_connection.execute("SELECT pg_sleep(2);")
# If the timeout worked, this statement will not execute.
# Currently, it does execute, which means my timeout isn't working.
print("Query successfully complete. Got result_proxy")
答案 0 :(得分:7)
您可以通过statement_timeout
设置options
parameter in libpq
等配置值。您可以psycopg2
作为connect
调用的一部分在connect
中访问此参数。您可以通过connect_args
参数将其他参数传递给SQLAlchemy的engine = create_engine(..., connect_args={"options": "-c statement_timeout=1000"})
调用。所以,把它们放在一起:
decay_rate = 0.99 # decay factor for RMSProp leaky sum of grad^2