使用MySQL限制每天发送的访问者

Question

I'm setting up a website that uses PHP and MySQLi. The website sells traffic to other website owners. Here's the structure of my table that monitors active campaigns:

id  url                     hits_ordered    hits_sent   status
1   http://google.com       1000            120         'Active'
2   http://facebook.com     3000            2421        'Inactive'
3   http://yahoo.com        400             12          'Active'

What I want to do is give people the option to have hits spread across 1-90 days, that way they don't just get flooded with them all at once. So, I was going to change the structure to this:

id  url                     hits_ordered    hits_sent   period      status
1   http://google.com       1000            120         25          'Active'
2   http://facebook.com     3000            2421        20          'Inactive'
3   http://yahoo.com        400             12          60          'Active'

The column period would represent the number of days the campaign should be spread across.

But I'm thinking that it may be better to add two columns instead, one that would show the time the campaign started ( start_time ) and when the campaign should be completed ( end_time ).

If I had start and end columns, would there be any way with MySQL to mathematically calculate if a campaign was behind schedule and select it?

For example, if someone order 10,000 hits and wanted them delivered over the course of 5 days. If on the 4th day, they only have 2000 hits, obviously it would be "behind schedule" because it should have between 8000-10000 hits.

I know that there's ways I could calculate this with PHP. I'm just trying to figure out if there would be a way to do this with purely MySQL to select all rows that are behind schedule.

It seems like there should be a way to query these rows since MySQL can perform math, knows the total hits sent, the total amount order, and the start/end dates.

================

Another example for @ICE below. This row would be behind schedule:

id  url                     hits_ordered    hits_sent   start_date              end_date
1   http://google.com       1000            5           2016-12-23 00:00:00     2016-12-28 00:00:00

It's behind because only 5 hits have been sent since it was ordered on the 23rd. It's now the 25th and ends on the 28th. So according to my logic above, it should be considered "behind schedule".

And this row would not be:

id  url                     hits_ordered    hits_sent   start_date              end_date
1   http://google.com       1000            999         2016-12-23 00:00:00     2016-12-28 00:00:00

In this example, the dates are the same, but 999 our of 1000 hits have been sent. It would not be behind schedule.

Answer 1

No need for column period . Create end_time and start_time as DATETIME . after that you can select them with datediff . It returns difference between two dates in days:

select * from TABLE where
datediff(end_time,start_time) > datediff(now(),start_time) and
datediff(now(),start_time)>3 and //just select rows with more than 3 days after start_time. This can help to have more logical result as you want
hits_sent + (datediff(end_time,now()) * (hits_send/datediff(now(),start_time))) + 20 < hits_ordered //means count hits happened + we guess this much hits can happen on remaining days based on the remaining days and hits_send + fluctuation < what they ordered

Explaining

hits per day until now

hits_send/datediff(now(),start_time)

remaining days

datediff(end_time,now())

With what happened in days before now we can guess this much hits could happen in remaining days

(datediff(end_time,now()) * (hits_send/datediff(now(),start_time))) + 20

+20 is the fluctuation. We guess maybe some day they can have 20 hits more.

Answer 2

You can compare the percentage of the remaining hits and the percentage of days of the campain, if the percentage of the hits_sent is above the percentage of the campain duration percentage, just disable it with a bool expresion.

select *,((hits_sent/hits_ordered)<datediff(now(),start_date)/datediff(end_date,start_date)) as active, (((hits_sent/hits_ordered))/(datediff(now(),start_date)/datediff(end_date,start_date)) ) as status from orders where 1 order by active desc

In a SQLFIDDLE gives the active or not campain, and give you a number that represent if the campain is on schedule (near 1) below schedule (near 0) and above 1, when its already complete