如何使用 Python 自动执行 SQL 查询中的日期

Question

我有 SQL 连接并命名为 db_connection，我每周运行一次

tran= pd.read_sql('SELECT order_id, p.user_id, p.barcode, title, o.timestamp, amount, p.name, qty FROM transactions t left join orders o on t.order_id = o.id left join products p on t.product_id = p.id where date(o.timestamp)<=\'2017-09-27\' and date(o.timestamp)>=\'2017-09-21\' and p.user_id = \'63165\' ', con=db_connection)

tran2= pd.read_sql('SELECT order_id, p.user_id, p.barcode, title, o.timestamp, amount, p.name, qty FROM transactions t left join orders o on t.order_id = o.id left join products p on t.product_id = p.id where date(o.timestamp)<=\'2017-09-27\' and date(o.timestamp)>=\'2017-09-21\' and p.user_id = \'62345\' ', con=db_connection)

...

tran22= pd.read_sql('SELECT order_id, p.user_id, p.barcode, title, o.timestamp, amount, p.name, qty FROM transactions t left join orders o on t.order_id = o.id left join products p on t.product_id = p.id where date(o.timestamp)<=\'2017-09-27\' and date(o.timestamp)>=\'2017-09-21\' and p.user_id = \'78345\' ', con=db_connection)

比如我想把日期全部改成date(o.timestamp)<=\'2017-10-04\' and date(o.timestamp)>=\'2017-09-28\'，但是这需要时间，而且容易出错。有没有办法让日期只被调用一次？

Answer 1

Python

虚拟解决方案是使用占位符来编辑单个条目：

start = # evaluate or hardcode date
end = # evaluate or hardcode date

tran = pd.read_sql('... where date(o.timestamp)<=\'{END}\' and date(o.timestamp)>=\'{START}\' ...'.format(START=start, END=end), con=db_connection)
...
tran1 = ...

熊猫的参数

start = # evaluate or hardcode date
end= # evaluate or hardcode date
tran = pd.read_sql('... where date(o.timestamp)<=%(end)s and date(o.timestamp)>=%(start)s ...',
               con=db_connection, params={"start":start, "end":end})

CURDATE

使用函数CURDATE，您可以对查询进行参数化。例如，如果您每周一运行查询，并希望提取上周（周一至周日）的数据，您可以设置条件：

WHERE DATE(o.timestamp)<=(CURDATE() - INTERVAL 1 DAY) and DATE(o.timestamp)>=(CURDATE() - INTERVAL 8 DAY)

这显然取决于您何时运行查询，但如果提取由调度程序自动执行，则效果很好。

或者，将它与DAYOFWEEK 结合起来，我们可以使它独立于它运行的星期几（再次：考虑上周一到周日））：

WHERE DATE(o.timestamp) >= (CURDATE() - INTERVAL (DAYOFWEEK(CURDATE())+5) DAY) AND DATE(o.timestamp) <= (CURDATE() - INTERVAL (DAYOFWEEK(CURDATE())-1) DAY)

显然，您必须根据需要调整日偏移量。