I am working with money expiration tracking problem at the moment (originally it is not money, but I have used it as a more convenient example).
An user can earn money from a platform for some mysterious reason and spent them for buying stuff (products, gifts etc.).
I am looking an algorithm (SQL query best case) to find a current balance of an user balance.
The events of spending and earning money are stored different database (MySQL) tables (let's say user_earned and user_spent). So in normal case, I would simply count user totals from user_earned and subtract spent money (total of user_spent).
BUT! There is a condition, that earned user money expires in 2 years if they are not used.
That means, if user have not used his money or used just a part of it, they will expire. If an user uses his money, they are used from the oldest not expired money record, so the balance (bonus) could be calculated in user's favor.
These are 5 scenarios with events in time, to have a better understanding on the case:
Both tables (user_earned and user_spent) have timestamps for date tracking.
I did something similar in one of my projects.
Looks like you need an additional table spends_covering with columns
spend_id, earhed_id, sum
So for each spends record you need to insert one or many rows into the spends_covering to mark 'used' money.
Then balance would be just sum of not used where date is less than 2 years.
select sum(sub.earned_sum-sub.spent_sum) as balance
from
(select e.sum as earned_sum, sum(sc.sum) as spent_sum
from earned e
left join spends_covering sc on e.earhed_id=sc.earhed_id
where e.date BETWEEN ...
group by e.earhed_id
having earned_sum > spent_sum) sub
It may be worth it to have two tables -- one (or more) with all the historical details, one with just the current balances for each 'user'. Be sure to use transactions to keep the two in sync.
Related
Basic question of language ability.
I am developing a database to keep track of market trades and provide useful metrics to the user. Most brokers do not supply enough information in the transaction .csv file which would be imported to this database to combine strategies and positions in a useful way, or in a way that I envision can be useful for users. For instance, combining a buy order on AAPL of 1000 shares that was filled in three separate transactions on the same order (1 order for 1000 shares filled initially with 200, then 350, then 450 shares). I could only think of assigning a trade group to each of these so that I can group them.
So, the example above, each transaction would be a separate record, I've created a column in my table with the alias of Trade Group, and each transaction would be assigned 1 under the Trade Group column. The sale of the 1,000 shares, no matter how many transactions it took to fill the order, would also be assigned to trade group 1.
My query combines the shares for both opening and closing transactions by using the trade group and transaction type (buy or sell). If there is a match, 1,000 shares on the buy side, and 1,000 shares on the sell side, then it runs some queries to provide useful data about the trade.
The problem I foresee is that the trade grouping can become cumbersome since it currently has to be manually input. I would like to develop a counter to automatically increment the trade group of each the opening and closing transactions every time the balance of shares = 0.
So if both the buy and sell of the above example belonged to trade group 1, and now I decide to open up a position of 2000 shares of AAPL, and subsequently sell them, those transactions would be automatically assigned trade group 2. And now that the balance of shares is 0 again, the next time I open and close a position on AAPL it will be assigned trade group 3.
That way, I don't need to clutter up my table with something that is manually input which can create mistakes. Instead, the query assigns the trade grouping every time it is run, and the necessary metrics supplied.
Is this something that can be done using SQL alone?
Thanks.
I am creating a database which need to allow booking a resource from start time to end time on a particular day. For example, I have 11 badminton courts. These courts can be booked for 1 hour and it can very also and in a day each court takes 18 bookings from morning 6 am till night 12 pm. (considering each booking is for one hour). The price of booking also varies from day to day, for example morning charges are more than day charges. Weekend charges are more than weekdays charges.
Now my question is, is it advisable to pre-populate slots and then book it for user depending on the availability. But in this case for the abobe example If I need to store slots for next 1 month then I will have to store 11*18*30 = 5940 records in advance without any real bookings.Every midnight I will need to run script to create slots. If no of clubs increases this number can become huge. Is this good design for such systems? If not then what is the better designs in these scenerios.
club name||court || date || start_time || end_time || status || charge ||
a c1 20/04/2015 6:00 7:00 available
a c1 20/04/2015 7:00 8:00 available
.
.
.
a c1 20/04/2015 11:00 24:00 available
.
.
a c11 20/04/2015 11:00 24:00 available
Now my question is, is it advisable to pre-populate slots and then book it for user depending on the availability. But in this case for the abobe example If I need to store slots for next 1 month then I will have to store 11x18x30 = 5940 records in advance without any real bookings.Every midnight I will need to run script to create slots. If no of clubs increases this number can become huge.
Yes. that is a horrible method. For the reasons you have stated, plus many more.
The storage of non-facts is absurd
The storage of masses of non-facts cannot be justified
If the need to write simple code is an issue, deal with that squarely, and elevate your coding skills, such that it isn't an issue (instead of downgrading the database to a primitive filing system, in order to meet your coding skills).
Notice that what you are suggesting is a calendar for each court (which is not unreasonable as a visualisation, or as a result set), in which most of the slots will be empty (available).
Is this good design for such systems?
No, it is horrible.
It is not a design. It is an implementation without a design.
If not then what is the better designs in these scenerios.
We use databases. And given its unequalled position, and your platform, specifically Relational Database.
We store only those Facts that you need, about the real world that you need to engage with. We need to get away from visualising the thing we need for the job we have to do (thousands of calendars, partially empty) and think of the data, as data, and only as data. Including all the rules and constraints.
Following that, the determination of Facts, or the absence of a Fact, is dead easy. I can give you the Relational Database that you will need, but you have to be able to write SQL code, in order to use the database effectively.
Data Model
Try this:
Resource Reservation Data Model
That is an IDEF1X data model. IDEF1X is the Standard for modelling Relational Databases. Please be advised that every little tick; notch; and mark; the crows foot; the solid vs dashed lines; the square vs round corners; means something very specific and important. Refer to the IDEF1X Notation. If you do not understand the Notation, you will not be able to understand or work the model.
I have included:
Storage of Facts (Reservations) only. The non-fact or absence of a Fact (Availability) is easy enough to determine.
club_resource_slot.duration in the Key to allow any duration, rather than assuming one hour, which may change. It is required in any case, because it delimits the time slot.
resource_code, rather than court number. This allows any club resource (as well as a court number) to be reserved, rather than only a badminton or squash court. You may have meeting rooms in the future.
Joel's reply re the rate table is quite correct in terms of answering that specific question. I have given a simpler form in the context of the rest of the model (less Normalised, easier to code).
If you would like the Predicates, please ask.
Code/General
You seem to have problems with some aspects of coding, which I will address first:
But the problem in this approach is if I need to find the availability of court based on game,location, date and time slot then I will have to load this rate table for all the clubs and the look into actual booking table if someone has already booked the slots. Is nt the better approach be if I keep the slots in advance and then someone book , jst change the status to booked. so That query will be performed entirely in DB without doing any computation in memory.
The existence of the rate table, or not, does not create an issue. That can be accomplished via a join. The steps described are not necessary.
Note that you do not need to "load this whole table" as a matter of course, but you may have to load one table or other in order to populate your drop-downs, etc.
When someone books a court, simply INSERT reservation VALUES ()
When someone cancels a reservation, simply DELETE reservation VALUES ()
Code/Data Model
Printing your matrix of Reserved slots should be obvious, it is simple.
Printing your matrix of Available or Available plus Reserved (your calendar visual) requires Projection. If you do not understand this technique, please read this Answer. Once you understand that, the code is as simple as [1].
You need to be able to code Subqueries and Derived tables.
Determination of whether a slot is Reserved or Available requires a simple query. I will supply some example code to assist you. "Game" isn't specified, and I will assume location means club.
IF (
SELECT COUNT(*) -- resources/courts reserved
FROM reservation
WHERE club_code = $club_code
AND date_time = $date_time
) = 0
THEN PRINT "All courts available"
ELSE IF (
SELECT COUNT(*) -- resources/courts that exist
FROM club_resource_slot
WHERE club_code = $club_code
AND date_time = $date_time
) = (
SELECT COUNT(*) -- resources/courts reserved
FROM reservation
WHERE club_code = $club_code
AND date_time = $date_time
)
THEN PRINT "All courts reserved"
ELSE PRINT "Some courts available"
Please feel free to comment or ask questions.
Assuming that each booking is for one hour (that is, if someone wants two hours on the court, they're taking two bookings of one hour each) it seems to me the most efficient storage mechanism would be a table Booking with columns Court, Date, and Hour (and additional columns for the person who booked, payment stated, etc..) You would insert one record each time a court was booked for an hour.
This table would be sparsely populated, in that there would only be records for the booked hourly units, not for the available ones. No data would be pre-generated; you would only create records when a booking occurred.
To produce an daily or weekly calendar your application would retrieve the booked hours from the database and join this with its knowledge of your hours (6am to midnight) to produce a visualization of court availability.
It is probably much more efficient from a data maintenance perspective to have a table with courts (1 record per court) and a table with bookings (1 record per booking).
The BOOKING record should have a foreign key to the COURT a booking start date/time and a booking end date/time. It would also have information about who made the booking, which could be a foreign key to a CUSTOMER table or it might be a fill-in name, etc., depending on how your business works.
Assume a simple database for hotel reservations with three tables.
Table 1: Reservations
This table contains a check-in and check-out date as well as a reference to one or more rooms and a coupon if applicable.
Table 2: Rooms
This table holds the data of all the hotel rooms with prices per night and number of beds.
Table 3: Coupons
This table holds the data of all the coupons.
Option #1:
If you want to get an overview of the reservations for a particular month with the total cost of each reservation, you'd have to fetch the reservations, the rooms for each reservation, and the coupon (if one is present).
With this data, you can then calculate the total amount for the reservation.
Option #2:
However, there is also another option, which is to store the total cost and discount in the reservation table so that it is much easier to fetch these calculations. The downside is that your data becomes much more dependent and much less flexible to work with. What I mean is that you have to manually update the total cost and discount of the reservation table every time you change a room or a coupon that is linked to a reservation.
What is generally recommended in terms of performance (option #2) version data independence (option #1).
UPDATE:
It is a MySQL database with over 500 000 rows (reservations) at this point, but is growing rapidly. I want to optimize database performance at an early stage to make sure that the UX remains fast and responsive.
Let me start to answer this with a story. (Somewhat simplified.)
2011-01-01 I reserve a room for two nights, 2011-03-01 and 2011-03-02. You don't tell me which room I'll get. (Because you don't know yet which room I'll get.) You tell me it will cost $40 per night. I have no coupons. You enter my reservation into your computer, even though you're already fully reserved for both those nights. In fact, you already have one person on the waiting list for both those nights. (Overbooking is a normal thing, not an abnormal thing.)
2011-01-15 You raise the rates for every room by $5.
2011-02-01 I call again to make sure you still have my reservation. You confirm that I have a reservation for two nights, 2011-03-01 and 2011-03-02, at $40. (Not $45, your current rate. That wasn't our deal. Our deal was $40 a night.)
2011-02-12 One person calls and cancels their reservation for 2011-03-01 and 2011-03-02. You still don't yet have a room you know for certain that I'll be able to check in to. The other person from the waiting list now has a room; I'm still on the waiting list.
2011-02-15 One person calls and cancels their reservation for 2011-03-01 and 2011-03-02. Now I have a room.
2011-03-01 I check in with a coupon.
You can store the "current" or "default" price with each room, or with each class of
rooms, but you need to store the price we agreed to with my
reservation.
Reservations don't reserve rooms; they reserve potential rooms. You
don't know who will leave early, who will leave late, who will
cancel, and so on. (Based on my experience, once in a while a room will
be sealed with crime scene tape. You don't know how long that will last, either.)
You can have more reservations than room-nights.
Coupons can presumably appear at any time before check out.
If you want to get an overview of the reservations for a particular
month with the total cost of each reservation, you'd have to fetch the
reservations, the rooms for each reservation, and the coupon (if one
is present).
I don't think so. The price you agreed to should be in the reservation itself. Specific rooms can't resonably be assigned until the last minute. If there's one coupon per reservation, that might need to be stored with the reservation, too.
The only reporting problem is in making sure your reports clearly report how much expected revenue should be ignored due to overbooking.
The response of your answer depends of the size of your database. For small database option #1 is better, but for huge database option #2 is better. So if you could say how many rows you got in table, and the database used (oracle, sqlserver etc.) you will have a more precise answer.
You can add a table holds the data of the rooms`s historical prices and reason for change.
Table 2 only records the latest price.
I am creating a website where members of the family can book a vacation in the family's sea cottage.
Each member is assigned 40 credits each year. A member must have enough credit points to book a vacation. (this credit system makes sure every member has equal opportunity to stay at the cottage)
However, not all weeks require the same amount of credit points, and last minute bookings require no credit points.
My question is: how should I store the credit points in the database? What is most efficient? Per day/per couple of days/per week? And how do I efficiently check how much credit points are needed for a given date range using mysql?
BTW: I am using the Zend Framework and I am familiar with the code etc., I just need some conceptual advice ;)
I don't know if I understand right the problematic here but :
You need to store in a "users" table the amount of credit remaining for each user.
You need to have another table with fields like "start period", "end period", "duration", "amount of points needed"
In the duration field, it could be an enum with values like "day", "week", "last_minute".
Then the logic will be in your PHP code, but with that, you'll be able to have the right number of credits needed whenever the user wants to use the cottage.
If I understand your case correctly, I think I would store something like:
number of remaining credits per user
a default number of credits per day
number of days before that counts as "last minute"
periods of deviating credit cost, with (start date, end date, cost per day)
Then, I would calculate the cost to book dynamically from this information whenever the user needs to make a booking. The last minute booking rule can be seen as a special deviating period where start date is now, end date is now+"number of days" and cost is zero.
I assume that no-one will book the cottage for a shorter duration than one day. If they will, I would change the use of "day" above to whatever unit of time would be sensible.
Just my thoughts. :-)
I apologize if this has been asked before, but I'm pretty new to this and unable to find an answer that addresses the situation I'm faced with.
I'm trying to put together a database to run behind our company website. The database will store information on customer invoices and payments. I'm trying to figure out if I should create a field for the invoice balance, or if I should just have it calculate when the customer account is accessed? I don't want to create redundant data, and don't want to have the chance that somehow the field wouldn't be updated, and would thus be incorrect...but I also don't want to create to large of a burden on the server - especially if we pull up an overview of customer accounts - which would need to then calculate the balance of every account. Right now we are starting from scratch, so I want to set it up right!
We are anticipating having a couple hundred customer accounts by the end of the year, but will most likely be up to a couple thousand by the end of next year. (Average number of invoices per customer would be roughly 2-3 per year.)
There are probably other things to consider as well. For example, what if your invoice consist of ID's of products in another table... and the prices of those other products change? When you go to generate the invoice, you'll have the wrong total in there for what the guy actually paid 6 months ago. So if its a situation like that, you'll probably want to store the total on the invoice. And I wouldn't worry too much about doing a little math if you go the other route, it's not likely to be a huge bottleneck.
Yes, remember that items/goods could and will change their prices over time. You need to have the invoice balance as of the day of the purchase. Calculating the balance on the fly could lead to wrong balances later on.
Invoice balance is essential data to store, however I think you meant account balance since you referred to that later.
Storing the account balance would be denormalizing it, and that's not how accounting databases are typically designed. Always calculate account balance from invoices minus payments. Denormalizing is almost always a bad idea, and if you need to optimize in the future, there are other places to cache data that are more efficient than the database.
In your use case, a query like that on a few thousand rows would be negligible anyway, so don't optimize before you have to.