Background:
We are setting up a promotions system to give away free products to registered customers. We're trying to design a database which is flexible enough to handle multiple products, and giveaways. The requirements are that products may be given away on a drip basis on a first come basis to qualified customers.
Example:
Apple wants to give away 1000 ipads in March.
They want to give away maximum of 1 per hour.
They want to give it to customers who are in California or New York.
They want to limit how many free ipads a customer can get (limit 1 per 15 days).
Data Structure:
Products - 1 entry per unique product. e.g. Apple iPad
ProductGiveAways
ProductID: AppleIpad
Quantity:1000
StartDate: 03/01/2014
End Date 03/31/2014
CustomerState: California,NewYork
PurchaseLimitDays: 15
Problem:
With the above structure we are able to do a query against our customers table and find out which are qualified for the promotion.
What I cannot figure out is the best way to:
Query customers in California or New York (is this a good use case for a join and another table?)
When a customer logs in to see what free items are not available to him, how can I exclude the Apple iPad if the customer has already gotten this freebie?
In other words:
Say amazon.com wants to show me DVDs which I have not already bought. What is the proper way to query that?
Is the right approach to first get a list of previously bought products and then Query with a NOT clause?
I'm assuming you'll have a table for what has been given away. In this table I would include a column for recipient id which can map back to the customer table. You can then create queries to find eligible recipients by searching for customers who have not met disqualifying conditions.
select customerid
from customer
where customerid not in (
select recipientid
from givenaway
where ..... and ....
)
Because there's not a definitive data structure defined, I'm going to use the following which you can tailor to whatever data structure you have designed yourself:
Product
ProductId - INTEGER (IDENTITY and PRIMARY KEY)
ProductName - VARCHAR
States
StateId - INTEGER (IDENTITY and PRIMARY KEY)
StateName - VARCHAR
Customer
CustomerId - INTEGER (IDENTITY and PRIMARY KEY)
StateId - INTEGER (FOREIGN KEY)
Promotion
PromotionId - INTEGER (IDENTITY and PRIMARY KEY)
ProductId - INTEGER (FOREIGN KEY)
Quantity - INTEGER
StartDate - DATETIME
End Date - DATETIME
PurchaseLimitDays - INTEGER
PromotionState
PromotionId - INTEGER (FOREIGN KEY)
StateId - INTEGER (FOREIGN KEY)
So in answer to your questions:
Query customers in California or New York (is this a good use case for a join and another table?)
Personally I would join to a centralized state table (PromotionState) in my above example, I'm sure there's a better way but you could do a condition such as:
WHERE
(SELECT COUNT * FROM PromotionState x WHERE x.PromotionId = p.PromotionId) = 0
OR NOT(ps.PromotionId IS NULL)
Alternatively you could do a GROUP BY and HAVING, using all the other columns as the items to GROUP BY and something like HAVING COUNT * = 0 OR HAVING SUM CASE WHEN (Conditions met) THEN 1 ELSE 0 END = 0
When a customer logs in to see what free items are not available to him, how can I exclude the Apple iPad if the customer has already gotten this freebie?
Say amazon.com wants to show me DVDs which I have not already bought. What is the proper way to query that?
As I've said you could use GROUP BY and HAVING to determine whether an item has been previously "won" by either using COUNT or SUM
Is the right approach to first get a list of previously bought products and then Query with a NOT clause?
There are probably better ways, sub queries can get very heavy and sluggish, I'd recommend trying some of the above techniques and then using a profiler to hopefully make it more efficient.
Some database design
First, when you set the CustomerState to California,NewYork you are violating the First Normal Form of database design.
So let's reorganize your domain model.
State - 1 Entry per unique state
...
Customer - 1 Entry per unique customer
StateId: (California|NewYork|...)
...
Product - 1 Entry per unique product
...
ProductGiveAways - Many entries per product
ProductID
Quantity
StartDate
End Date
PurchaseLimitDays
...
ProductGiveAways_State
ProductGiveAwaysId
StateId
...
Customer_Product - 1 Entry per bought product by customer
CustomerId
ProductId
PurchaseDate
...
Technical issues
When you want to query custoners in California or New York, all you have to do now is :
// This is just an example, you have to change the 'California', 'New York' with their ids
SELECT * FROM Customer WHERE StateId IN ('California', 'New York')
When a customer logs in to see what free items are available to him :
// It's not an accurate sql, just an example
SELECT Product.*
FROM Product
JOIN ProductGiveAways ON ProductId
JOIN ProductGiveAways_State ON ProductGiveAwaysId
WHERE ProductId NOT IN (
SELECT ProductId FROM Customer_Product JOIN ProductGiveAways ON ProductId
WHERE CustomerId = /* the customer id */
AND ( TO_DAYS(now()) - TO_DAYS(PurchaseDate) ) < PurchaseLimitDays
)
AND StateId = /* customer StateId */
AND StartDate < now() < End Date // Elligible ProductGiveAways
For Laravel We Use Something Like this, i hope you can relate to this query or you can use online laravel query converter for using it in mysql ( orator )
$user_id = auth()->user()->id;
Product::where('status', 'active')->whereNotIn('id', function($query) use ($user_id) { $query->select('product_id')->from(new OrderProduct->getTable())->where('user_id', $user_id)->where('status', 'delivered'); });
Related
I have to create a database for a tool shop and since I am not able to get MySQL running, because of my computer, I wanted to ask for help with my queries.
I have several entities, like customer, car, repair, repair team member, employee, service, tire storage, personal file, sales.
Now I have to do following queries:
Sort every car that contains the combination "123" inside the vehicle identification number. (Result column: VehicleID)
Here I would say SELECT * FROM Vehicle WHERE VIN = '%123%';
A list of every sales in 2021. (Result column: document number)
Here I don't know how to connect both entities (service, sales) and filter by year.
The total revenue generated from sales in 2020
An overview of all the sales to the customer with the customer ID "4711" (Result column: Date of sale, product number, price)
What about -> SELECT * FROM Verkauf WHERE Verkauf.Dienstleistung_D_ID IN ( SELECT D_ID FROM Dienstleistung WHERE Kunde_K_ID == "K4711" )
Overview of all customers who have been sold something by an employee with employment year before 2010. The output should be sorted by last name and first name of the customer. (Result columns: Customer number, first name, last name, personnel number, last name of employee, year of employment).
I hope you can help me, since I am not able to create the code, even with SQLite, MySQL etc. so I am not able to test my queries and hope that you can help me.
ER Diagram
My ER diagram is in German, I hope you understand the structure.
Reparatur - repair,
Fahrzeug - car,
Kunde - customer ,
Dienstleistung - service,
Verkauf - sales ,
Reifenlagerung - tire storage,
Reparaturteammitglied - repair team member,
Mitarbeiter - employee,
Personalakte - personal file,
My reservation system allows us to purchase credits for clients in terms of pre defined packages. I'm struggling with how I record and calculate available credits.
Let's say we're talking about a car wash service. A client can have multiple cars and can purchase the following services, 'Wash and Wax' and 'Detailing'.
Client 1 has two cars, Car A and Car B. He brings them both in and purchases:
Car A - 1 Wash and Wax
Car A - 1 Detailing
Car B - 10 Wash and Wax
Car B - 1 Detailing
This generates 4 rows in my Purchases table, one for each service purchased.
In my DB I have two related tables tracking purchases and reservations. Table 1 Purchases, Table 2 Reservations.
In Purchases I have the following fields of note:
id
client_id
car_id
service_id
credits_purchased
credits_scheduled
credits_used
cart_id
Then in my Reservation table I have the following fields of note:
id
client_id
car_id
service_id
reservation_date
completed_datetime
car_in_datetime
car_out_datetime
purchase_id
I track the credits available by updating the Purchases table fields credits_used and credits_on_schedule as events happen.
For example, when the client makes a reservation the system adds a new record in the Reservations table, once this happens the system also runs an update query and adds +1 to the related Purchases table credits_on_schedule. When the Reservation is updated to complete the system also updates the Purchases table and adds -1 to credits_on_schedule and +1 to credits used. Simple math between credits_purchased, credits_used, and credits_on_schedule derive what credits are available for a client to use.
I feel like this isn't a good way to track the credits. My question is what is a better implementation? Should I just track credits_purchased then use count queries on the Reservation table to calculate credits_used and credits_on_schedule? Should I be using a pivot table to track? I can't seem to wrap my head around what is the cleanest design.
It looks to me that the design is ok in general.
A reservation can only have one purchased related to it so purchase_id field is a foreign key in Reservation table.
Nevertheless, my advise to you is to create a log system of all these record updates.
As you mentioned, as events are fired the system updates the calculated fields.
What if for some reason the system fails at a certain point? You should be able to track these events.
One way to avoid this is, as you mentioned, calculate credit_used by a count query on all completed reservations.
I have two tables in MYSQL, one contain a list of options a customer can pick from and the other has the cost of the options. I would like a view that returns the cost of each customers option set. For my example I'll use the analogy of buying a new car.
Customers table
customer - sunroof - mag_wheels - spoiler
--------------------------------------------------
John - true - true - false
Steve - false - true - false
Lucy - false - false - false
Options table
option - price
-----------------
sunroof - 100
mag_wheels - 150
spoiler - 75
Desired results
customer - cost
-----------------
John - 250
Steve - 150
Lucy - 0
or this would do, as I good easily multiply selected by price and then group by customer
customer - option - selected - price
------------------------------------------
John - sunroof - true - 100
John - mag_wheels - true - 150
John - spoiler - false - 75
Steve - sunroof - false - 100
Steve - mag_wheels - true - 150
Steve - spoiler - false - 75
Lucy - sunroof - false - 100
Lucy - mag_wheels - false - 150
Lucy - spoiler - false - 75
I've been puzzling over this for hours now and I can't even figure out where to start, a join seems out of the question as there are no common elements to match. I wonder if using UNION is the answer but I can't figure out how to combine row values with column headings.
If anyone could point me in the right direction I'd be ever so grateful, double points if you come up with a solution that dynamically picks up the different options so I could add more in the future without rewriting the query.
Many thanks in advance.
The reason I wanted to have all the options as a single row is I was hoping to use Access to make a form for picking the options and I couldn't figure out how a single form could create multiple rows.
This is a horrible data layout. You should have an association table, with one row per customer and option.
But, you can do it:
select c.customer, sum(o.cost) as cost
from customers c left outer join
options o
on (c.sunroof = true and o.option = 'sunroof' or
c.mag_wheels = true and o.option = 'mag_wheels' or
c.spoiler = true and o.option = 'spoiler'
)
group by c.customer;
EDIT:
You do not want all options in a single record. Instead, you need an association table:
create table customer_options (
customer_optionid unsigned auto_increment,
customer varchar(255) references customer(name),
option varchar(255) references option(option)
);
Actually you should really have integer primary keys for all the tables, and use them for the foreign key references. If you need data in the output in the question, then just write a query to return it in that format.
Looking at the table structure, even I think it will not be possible to write joins because as you mentioned, the table structure doesn't have a relation between them.
I am assume you have just started the project, so it's time you first re-visit your DB structure and correct it.
Ideally, you should have a customer table with a customer id.
Then you should have products table with product id.
One table which will have data on what products customers have purchased - something like customer_products. This will be a one to many relation. So customer 1 can have product 1,3 and 5. Which would means in customer_product there will be three entries.
And then when you want to do a sum total, you can first join the table customer, product based on customer_product and then do a sum of the price also to get the total amount for individual customer.
Bad design. You must a have a customers table like this:
custumer_id
customer_name
other fileds...
On the other hand you should have an accesories table, where you usually describe each item-
accesory_id
accesory_name
supplier_id
country_of_origin
other stuff
Also an accesory_price table, where prices are added due the fact that prices change.
accesory_id
price
active_price
date_price_added
An finally you should relate all in a customer_accesory table:
customer_id
accesory_id
By having this, you can join tables and select both customer basket size and customer preferences of accesories. Basket size, or the amount purchased by each customer can be summarized SUM, AVG, COUNT or you can pivot data using GROUP_CONCAT in order to generate high quality reports.
For those interested in the reasoning behind this question: I have an e-commerce site that works fine, but has no gift certificate capabilities. Adding monetary GCs should be pretty simple, but I'd also like to allow the gifting of specific products (sounds odd but is relevant to my industry). So I plan to create a new table to house gift certificates that are linked to a specific user and product, and I need an efficient way to evaluate that table on the cart and checkout pages.
Imagine tables exist that look similar to the following:
CartContents
CartID Integer (Unique sequential row identifier)
UserID Integer
ProductID Integer
Quantity Integer
Gifts
GiftID Integer (Unique sequential row identifier)
ProductID Integer
UserID Integer
Quantity Integer
This is an overly simplified layout, but demonstrates the idea. The first table lists items in the user's cart; one record per product (though real products will have additional details that may vary). The product table has further attributes on products but I don't list it here for simplicity. The second table is a set of gift certificates, each for a specific product, that have been presented to this user ID.
Table data may look like the following:
CartContents
CartID UserID ProductID Quantity
1 1 1 1
2 1 2 2
3 1 1 2
4 2 3 1
Gifts
ProductID UserID Quantity
1 1 1
2 1 1
3 3 1
Is it possible to construct a single query that provides one row per cart item and links the above two tables taking into account that each gift may only link to each cart item once? Or does this need to be handled in a script?
In other words, because user 1 has product 1 in their cart twice, and they have only been promised one free product 1, the query should return a matching Gifts record for cartID 1, but not cartID 3. The query, pulling for user ID 1, would return:
CartID ProductID Quantity unpaidQuantity
1 1 1 0
2 2 2 1
3 1 2 2
Or
CartID ProductID Quantity unpaidQuantity
1 1 1 1
2 2 2 1
3 1 2 1
I realize that the fact that there is more than one 'right' answer to this question raises a red flag. In reality it doesn't matter which cart record each GC is applied to, as the end result (the price) will work out the same. I'm perfectly happy to say the 'first' (lowest cartID) is the one that should be linked.
My assumption is that the database will be far more efficient at this than any script I could write; I'd even be willing to bet there's some crazy type of join I've never heard of specifically designed for it. I am also assuming that any such ColdFusion script may be somewhat complicated and thus take a fair amount of development and testing time while a single query may be relatively simple (though apparently beyond my limited SQL capabilities). If I'm incorrect in this I'd appreciate any thoughts on that as well.
My setup, if it matters:
MySQL 5.0
ColdFusion 9
Windows 2000 AS
Edit:
It sounds like the quantity column is really going to cause issues, so let's continue assuming that quantity does not exist on the Gifts table. It still must exist on cartContents, however.
I thought of another way of doing this that just requires and additional group by and join. However, it requires a unique id on CartContents. I'm not sure i this is what CartId is supposed to be. However, it seems that a user could have more than one cart, so I assume not.
The idea is to identify the first record for a given product in each cart. Then, use this information when joining in the gifts.
select CartID, UserID, ProductID, Quantity, FirstCCId
from CartContents cc join
(select CartID, UserID, ProductID, min(CartContentsId) as FirstCCId
from CartContents cc
group by CartID, UserID, ProductID
) ccmin
on cc.CartId = ccmin.CartId and cc.UserId = ccmin.UserId and
cc.ProductId = ccmin.ProductId left outer join
Gifts g
on cc.ProductID= g.ProductId and cc.UserID = g.userId and
cc.CartContentsId = ccmin.FirstCCId
This works when the gift is applied to only one product line row. If the quantity for the gift is actually larger than the quantity on any given line, this query still only puts it on one line.
Does this work?
select c.cartid, c.productid, c.quantity, c.quantity -
case
when (select sum(c2.quantity) from CartContents c2
where c.userid = c2.userid
and c.productid = c2.productid
and c.cartid < c2.cartid) <
(select g.quantity from gifts g
where c.userid = g.userid
and c.productid = g.productid) then
(select g.quantity from gifts g
where c.userid = g.userid
and c.productid = g.productid) -
(select sum(c2.quantity) from CartContents c2
where c.userid = c2.userid
and c.productid = c2.productid
and c.cartid < c2.cartid)
else 0
end UnpaidQuantity
from CartContents c
where userid = 1
I have a custom shop, and I need to redo the shipping. However, that is sometimes later, and in the meantime, I need to add a shipping option for when a cart only contains a certain range of products.
SO there is a ship_method table
id menuname name zone maxweight
1 UK Standard ukfirst 1 2000
2 UK Economy uksecond 1 750
3 Worldwide Air world_air 4 2000
To this I have added another column prod_restrict which is 0 for the existing ones, and 1 for the restricted ones, and a new table called ship_prod_restrict which contains two columns, ship_method_id and item_id, listing what products are allowed in a shipping category.
So all I need to do is look in my transactions, and for each cart, just check which shipping methods are either prod_restrict of 0 or have 1 and have no products in the cart that aren't in the restriction table.
Unfortunately it seems that because you can't values from an outer query to an inner one, I can't find a neat way of doing it. (edited to show the full query due to comments below)
select ship_method.* from ship_method, ship_prod_restrict where
ship_method.`zone` = 1 and prod_restrict='0' or
(
prod_restrict='1'
and ship_method.id = ship_prod_restrict.ship_method_id
and (
select count(*) from (
select transactions.item from transactions
LEFT JOIN ship_prod_restrict
on ship_prod_restrict.item_id = transactions.item
and ship_prod_restrict.ship_method_id=XXXXX
where transactions.session='shoppingcartsessionid'
and item_id is null
) as non_permitted_items < 1 )
group by ship_method.id
gives you a list of whether the section matches or not, and works as an inner query but I can't get that ship_method_id in there (at XXXXX).
Is there a simple way of doing this, or am I going about it the wrong way? I can't currently change the primary shipping table, as this is already in place for now, but the other bits can change. I could also do it within PHP but you know, that seems like cheating!
Not sure how the count is important, but this might be a bit lighter - hard to tell without a full table schema dump:
SELECT COUNT(t.item) FROM transactions t
INNER JOIN ship_prod_restrict r
ON r.item_id = t.item
WHERE t.session = 'foo'
AND r.ship_method_id IN (**restricted, id's, here**)