Question

Hi i am totally confused with a date logic in my mysql query for a cron job to be run everyday at 12:00 AM

I am working on a auto listing website where the car listings are having a expiry date in mysql datetime format.

All the expired listings will be deleted from the website after 7 days from the datetime of the expiry

When the cron job will run it has do following things

Task 1 – Send an email alert to the users telling them that their listing has expired.

So I need to select all those listings which have expired since last time the cron job has been run and not include listings before that in order to send the expiry alert email only once per listing.

I tried following sql query for this task (Again confused with this as well)

SELECT car_id FROM cars WHERE expiry_date > DATE_SUB(NOW(), INTERVAL 1 DAY) AND expiry_date < NOW()

Task 2 – Will send an email alert to users telling them that listing is going to be permanently deleted after 24 hours.

So I need to select all those listings which are going to be deleted in more than 24 hours / 6 days have passed since they were expired and i need to make sure that they get minimum 24 hours time to renew them. Also i need to select / build the sql query in such a way that only those listings get selected which are going to expiry in 1 days and not other in order to avoid multiple email alerts instead of one time email alert

I tried following sql query for this task (I am totally confused with this query)

SELECT car_id FROM cars WHERE expiry_date > DATE_SUB(NOW(), INTERVAL 7 DAY) AND expiry_date < DATE_SUB(NOW(), INTERVAL 1 DAY)

Task 3 – Delete all the listings which were expired more than 7 days ago

I tried following sql query for this task

SELECT car_id FROM cars WHERE expiry_date < DATE_SUB(NOW(), INTERVAL 7 DAY)

Please help me in perfecting all the 3 queries so that the cron does it job exactly as i want. Also please let me where it has to >= (greater than or equal to) or <= (less than or equal to)

Here is the sqlfiddle table structure and couple of records (though they are not expired yet)

http://sqlfiddle.com/#!2/cfcdf

I will really appreciate the help.

Answer 1

There is nothing significantly wrong with your queries, if you do not understand the functions that you have used then google them and read about them until you do.

There is a fundamental problem in your approach in that it relies on the cron job being run at exactly 24 hour intevals – to the milisecond – or there will be double ups and/or omissions.

You need another table to store details of when your batch program last ran; intitialise this with 1 row with a date a long time in the past so that we have a starting point.

You can get the most recent batch by SELECT MAX(date_ran) FROM BatchRecordTables. Store this in a local variable T0. Get the current time, store this in a local variable T1 (Do not use NOW() in multiple queries as they will be slightly differant times and you need them to be the same). I do not know the syntax for this is MySQL – you will have to look it up.

Your situations then become.

1. Send email to people whose listings have expired since that last time the cron job was run i.e. SELECT car_id FROM cars WHERE Expiry_Date BETWEEN T0 AND T1. This will only select people whoose listings have expired between this batch and the previous one.
2. For the second case, we need to know that these people have got the first email i.e. that their listing had expired before the last batch run so SELECT car_id FROM cars WHERE Expiry_Date BETWEEN DATE_SUB(T1, INTERVAL 6 DAY) AND T0. This will only select people whoose listings expired before the last batch (i.e. they got the exprired email) and more than 6 days ago.
3. Same logi applies – we want to know they got the second email. SELECT car_id FROM cars WHERE Expiry_Date BETWEEN DATE_SUB(T1, INTERVAL 7 DAY) AND DATE_SUB(T0, INTERVAL 6 DAY)

May I also suggest that you do not permenantly delete the listings but either copy them to a DeletedListings table or Flag them with a Deleted column – each has its own pros and cons. In the information age, never throw data away – you never know when it might be valuable.