求一段Flask(SQLAlchemy)的代码优化，关于在一个循环里重复select的 .

Question

#计算7天内销量，并把每件商品的不同尺码库存分别显示
    day7=today-datetime.timedelta(days=7)
    logs7=db.session.execute("SELECT product_name,goods_id,sum(number) as sellsum FROM sdb_sell_logs where createtime>:createtime group by goods_id order by sum(number) desc limit 0,20",{"createtime":datetime_timestamp(day7.strftime("%Y-%m-%d 00:00:00"))}).fetchall()
    goods_store7={}
    for item in logs7:
        store=db.session.execute("SELECT pdt_desc,store from sdb_products where goods_id=:goods_id",{"goods_id":item.goods_id}).fetchall()
        item_store=[]
        for store_item in store:
            item_store.append([store_item.pdt_desc,store_item.store])
        goods_store7[item.goods_id]=item_store

这段代码要怎么优化请问？

感觉这个for item in logs7:的循环里要处理N次的select效率好低下。

Answer 1

你可以这样处理，做成一个查询（空间换取时间）

product_name, goods_idA, sum(number), pdt_descA, storeA
product_name, goods_idA, sum(number), pdt_descB, storeB
product_name, goods_idA, sum(number), pdt_descC, storeC
product_name, goods_idB, sum(number), pdt_descA, storeA
product_name, goods_idB, sum(number), pdt_descB, storeB
product_name, goods_idB, sum(number), pdt_descC, storeC

两个SQL做个连接，剩下的就是输出时去控制成你想要的结构就好了。