MYSQL stock gap prediction - mysql

Question

I've been scratching my head on a way to get this working for quite some time now, and as it's so specific I can't seem to find the answer elsewhere on StackOverflow. Any answers or suggestions would be appreciated.
I'm working to produce a report on that will inform a user of stock management and consumption. Calculating this with the values of current_stock and daily_stock_consumption is easy... But when we add shipments incoming into the process things become very complicated. Getting MYSQL to return the first day that each stock reaches 0 is leaving me blank.
For instance, let's say we are selling fruit and we have three tables fruit, fruit_sold and fruit_orders
fruit
|---------------------|------------------|
| id | fruit |
|---------------------|------------------|
| 1 | Apples |
|---------------------|------------------|
| 2 | Oranges |
|---------------------|------------------|
| 3 | pears |
|---------------------|------------------|
fruit_sold
|---------------------|------------------|------------------|------------------|
| id | fruit_id | sold_per_day | day_sold |
|---------------------|------------------|------------------|------------------|
| 1 | 1 | 101 | 1585695600 |
|---------------------|------------------|------------------|------------------|
| 2 | 2 | 445 | 1585695600 |
|---------------------|------------------|------------------|------------------|
| 3 | 3 | 214 | 1585782000 |
|---------------------|------------------|------------------|------------------|
| 4 | 1 | 512 | 1585782000 |
|---------------------|------------------|------------------|------------------|
fruit_orders
|---------------------|------------------|------------------|
| id | arriving | fruit_id |
|---------------------|------------------|------------------|
| 1 | 1592607600 | 1 |
|---------------------|------------------|------------------|
| 2 | 1586905200 | 3 |
|---------------------|------------------|------------------|
| 3 | 1590534000 | 2 |
|---------------------|------------------|------------------|
| 4 | 1588287600 | 3 |
|---------------------|------------------|------------------|
| 5 | 1593126000 | 1 |
|---------------------|------------------|------------------|
| 6 | 1592607600 | 2 |
|---------------------|------------------|------------------|
So far I have
SELECT SUM(fruit_sold.sold_per_day) / 2 AS consumption, fruit.fruit
FROM fruit_sold
LEFT JOIN fruit
ON fruit_sold.fruit_id = fruit.id
GROUP BY fruit_sold.fruit_id
How would I also return the number of days until the stock runs out (taking into account arriving orders and their arrival dates) for each fruit?
We want to return the following:
|---------------------|------------------|------------------|------------------|
| id | fruit | consumption | days_until_gone | |
|---------------------|------------------|------------------|------------------|
| 1 | apples | 306.5 | 2 |
|---------------------|------------------|------------------|------------------|
| 2 | oranges | 222.5 | 16 |
|---------------------|------------------|------------------|------------------|
| 3 | pears | 107 | 3 |
|---------------------|------------------|------------------|------------------|

Answer 1

You can get the numbers per day using union all, aggregation, and window functions:
select fruit_id, day_sold, sum(sold_per_day), sum(orders),
sum(sum(sold_per_day - orders)) over (partition by fruit_id order by day_sold) as net
from ((select fruit_id, day_sold, sold_per_day, 0 as orders
from fruit_sold
) union all
(select fruit_id, as day_sold, 0 as sold_per_day, count(*) as orders
from fruit_sold
group by fruit_id, day_sold
)
) f
group by fruit_id, day_sold;
I'm not 100% how this applies to your sample data, but it seems to basically provide the information that you want.