I have a database (for a pet sitting company) containing tables for the following:
Customers
emergency contacts
phone numbers
phone types
The phone numbers are stored in a separate table to allow for efficient storage of a virtually unlimited number of phone numbers per customer. The phone numbers table stores both the customer ID and the phone type ID in addition to the primary key. My question is - is the best way to allow for emergency contacts to have the same functionality with phone number records to add another field to the phone numbers table "emergency contact ID"? Or should I be storing emergency contacts in the same table as customer (and rename it Individuals)? If so, please tell me how to create a relationship between records in the same table.
Thanks so much,
Jessica
You have to ask yourself the question: how far should you really go to respect data normalisation rules?
I'm not sure what the PhoneTypes table would contain, but if it is a list of things like Mobile, Work, Home, iPhone you are probably going a bit too far already: you're not building a contacts application, you're building a Pet Sitting application, there are probably more important areas of the software that would demand your development time.
Increasing the complexity of software is costly: time to mimplement features tends to creep up, as well as complexity and along with it the risks of errors, cost of maintenance, and more often than not, performance suffers as well.
These contacts details are really just properties of the Customer.
A Customer can have multiple telephone numbers and multiple emergency contacts.
Usually these should be listed in order of importance, so if the needs arises, you call the most relevant person first.
Without more information, the way I would handle this would simply to leave 2 memo fields in my Customer table where the user of the app can enter that data in any way she pleases, so she can list it in the right order, make annotations as necessary (Call on Mondays only, Customer's mum, call after 11:00am only, etc).
You can further constrain data input if you wish, like have a textbox where the user enters the details before clicking an 'Add' button that will append the data to the field, for instance by using a semi-conlon or simply CrLf to separate records. The data can then be split on the semi-colon or CrLf and shown in a listbox on the form for better presentation.
You can handle both Customer Phone numbers and Emergency contact numbers in the same way.
This makes things simple: all the Customer data is in one table instead of being split across multiple tables, with no unnecessary joins, it won't take more space than using multiple tables (actually, it will save space). It makes reporting easy (you can simply show the customer list and that will show all available phone numbers for all customer without you having to do anything fancy), it makes searching easy as well.
Having multiple values in a single field is quite common for peripheral data.
Unless you absolutely need to separate contacts, and make complex reports based on them or make sure you can re-use them, you do not have to create tables for every bit of information. Let the application user enter what is relevant for the Customer.
Constrain data entry to format it and check its consistency if you want, but ultimately, unless the purpose of the software is to maintain a complex contact list, don't make it harder than it probably should. A bit of VBA and some string manipulation is sufficient to constrain the data, allow it to be rearanged in the order that's most relevant by the user, and it will make your app snappier by avoiding some complexity.
Anyway, I would start with something simple anyway and see later if splitting the data accross multiple tables makes sense.
Avoid premature optimisation.
However, if you feel you really need to handle this by the book, I would probably handle it as follow:
Store everything in a Contact table that could have properties like these:
ID: unique contact ID
PhoneNumber: TEXT
PhoneTypeID: (whatever that is if it links to your PhoneType table)
IsEmergencyContact: BOOLEAN
ContactName: TEXT, freeform, how to address the contact person
CustomerID: foreign key linking to the Customer table
Notes: MEMO, any useful info about the contact
Rank: INTEGER, a sortable rank of importance for this contact
If you want to decouple the Customer from the Contact, so you can re-use the contact for multiple customers, you would need an intermediary table:
The Contact table would become:
ID: unique contact ID
PhoneNumber: TEXT
PhoneTypeID: (whatever that is if it links to your PhoneType table)
ContactName: TEXT, freeform, how to address the contact person
Notes: MEMO, any useful info about the contact
And the CustomerContact table (that makes the many-to-many relationship possible):
CustomerID: foreign key linking to the Customer table
ContactID: foreign key linking to The Contact table
IsEmergencyContact: BOOLEAN
Rank: INTEGER, a sortable rank of importance for this contact
To display and manage the list of Contacts and list of Emergency contacts, you simply need to filter each listbox or subform where you show the information based on whether IsEmergency is true or false.
Now, if you want the same contact to have multiple phone numbers, you will have to split everything even further:
The Contact table would become:
ID: unique contact ID
ContactName: TEXT, freeform, how to address the contact person
Notes: MEMO, any useful info about the contact
A PhoneNumber table would contain:
ID: Phone record ID
ContactID: foreign key linking to The Contact table
PhoneNumber: TEXT
PhoneTypeID: (whatever that is if it links to your PhoneType table)
Notes: MEMO, any useful info about this particular phone number
Now you have 4 tables to store all the info you need and share it any way you want, so a Customer can have multiple contacts (emergency or not), Contacts can have multiple phone numbers, Contacts can be shared accross customers (so one customer's contact is another customer's emergency contact):
Customer
Contact
PhoneNumber
CustomerContact
As I said, doing it the right way will entail a lot more complexity than maybe you really need.
Be careful about not building complexity prematurely. It's nice to anticipate the worse case scenario, but often, it means that you are prematurely optimising and therefore spending time on an area of the software that is not as important as the core of your app.
You always have to ask yourself: should I spend 2 days implementing this or spend 2 days refining the UI, testing or adding code to ensure data integrity, etc?
More often than not, YAGNI
I'd store it the same way as you store phone numbers; in its own table. This allows you the ability to store multiple numbers, and some people may have multiple emergency contact numbers. You always want to think about scalability when designing a database, and plan for the most complex situations. For example, I would imagine with pet sitting that a lot of your customers would come through word-of-mouth, and it's very possible you'll use the same contact for multiple clients.
Your first instinct (storing customers and contacts in a single table) was correct. If you think about it, customers and contacts are both people. It's just that both customers and emergency contacts are a specialised case of people. We can model this using a relational DB.
Let's create a table to hold info about people:
create table tblPeople (
ID autoincrement primary key
, FirstName varchar(100)
, LastName varchar(100)
, Notes memo
)
Now let's have a table to hold info about customers, but enforcing the fact that customers must also be people:
create table tblCustomers (
ID long primary key
constraint Customers_ID
references tblPeople (ID)
, EmergencyContactID long
constraint Customers_EmergencyContactID
references tblPeople (ID)
)
This is called a one-to-one relationship and is used to implement specialisation--like inheritance in object-oriented programming.
You have a choice here. Do you want to let each person have an arbitrary number of phone numbers of arbitrary types? This is obviously more general and more powerful. But also more complicated. Or do you want to go back and just store a fixed number of phone numbers for each person?
Let's say you want to do the former just to take it all the way. In that case, you need a table to hold phone numbers:
create table tblPhoneNumbers (
ID autoincrement primary key
, PhoneNumber varchar(15)
)
Notice how we don't specify here anything about what type of phone number it is. That part is next:
create table tblPhoneNumberTypes (
ID autoincrement primary key
, PhoneNumberType varchar(20) not null
)
Now we associate each person with a phone number and type:
create table tblPeople_to_PhoneNumberTypes_to_PhoneNumbers (
PersonID long not null
references tblPeople (ID)
, PhoneNumberTypeID long not null
references tblPhoneNumberTypes (ID)
, PhoneNumberID long not null
references tblPhoneNumbers (ID)
, constraint People_to_PhoneNumberTypes_to_PhoneNumbers_PK
primary key (
PersonID
, PhoneNumberTypeID
, PhoneNumberID
)
)
Here each person (and therefore each customer and each emergency contact) can have an arbitrary number of phone numbers of arbitrary types. Hence this is actually a many-to-many-to-many link table. I believe that is the key (or let's say 'secret sauce') to your contact-phone type-phone number model.
In link tables like the above I prefer to use multiple-column primary keys as I feel there is no useful purpose to an integer primary key column. Here the primary key enforces the fact that each person-and-phone-number combination should be listed only once, with one phone number type.
Note that the above is all valid Access ANSI-92 SQL.
Related
I am creating a web application for internal use. One of the main features is the ability for the users to enter Sales Call Reports. My question is simply: Do I store the Employee Name and Company Name in the Call Reports table or do I only store the employee number and customer number in the table, as foreign keys and then retrieve the Employee Name and Customer Name, using the foreign keys, for display purposes? It seems like the foreign keys option would be procedural correct, however, it would be much easier to retrieve data from only 1 table when displaying the call reports.
Usually you'll save future-you a lot of pain by storing them separately as you've described.
On the face of it, and initially it seems like lots of extra work, but you'll end up with a much more flexible system, and 2 JOINS in a query isn't a lot.
I am creating the database structure of an ecommerce with Mysql and INNODB engine.
Point 1: To create multiple addresses for the customers i have this tables
Am I doing it in the correct way? And how should I store the default address (in which table)?
Point 2: I have another table called "Suppliers", should i just connect it to addresses with a "supplier_address" table or is there a better way?
Point 3: What about the tables cities and countries? Should i add something or is that ok? Maybe a field "district" in another table beetween the two?
In my view you're making this far too complex. There's no need to make your address schema so over-normalized. Most systems I've seen that handle multiple customer addresses have a customer table like yours, and then have an address table, as follows:
customer_id
address_ordinal (small number for each customer: 0,1,2,3 etc).
primary (boolean)
address_1
address_2
locality (city, village, etc)
province (state, etc)
postcode (zip, postcode etc)
country
customer_id is a foreign key to the customer table. The primary key is a composite of (customer_id, address_ordinal). The primary column is true if the address is the primary one.
Regarding your question about suppliers, you might want to create a common table called "contacts", and give both your customers and suppliers contact_ids.
If your system contains a reference table (perhaps something you purchase from a data supplier) containing (postcode, locality, province) rows, you can use that to help populate your address table. But you should avoid forcing your addresses to only contain hard-coded postcodes: those reference tables get out of date very fast.
I'll start out my answer with the ole cliche: "There's more than one way to skin a cat." That said, I have a few suggestions:
Point 1 - Assuming a customer can have multiple addresses (i.e. billing and mailing), then yes, you have the right idea in terms of the separate mapping table. As for adding a field to customer_addresses called default or preferred, or something like that, it's not a bad idea, but another option is to add a new field called address_type that would reference a separate table with two records, "Billing" and "Mailing" and/or whatever else you would want. Then, in whatever application you are coding that is going to use the address data, depending on what context i.e. if you're on the billing info page, then code the address type that you use on the page itself something like SELECT * FROM customer_addresses WHERE address_type = 2 /* Billing */.
Point 2 - Same as for customers.
Point 3 - Do you want to be able to display shortened country names? For example, abbreviate "United States" to "US," "Canada" to "CAN," or "United Kingdom" to "UK?" I'd consider adding a field for abbreviated country names for that purpose.
I'm trying to determine the best way to represent the concept of "linked" contacts (people) in an existing mysql database. Let me explain what that means:
I have a table with different contacts, each with an ID. I have another table with different time periods that a contact can belong to. Each period also have an ID. There's a N:1 relationship between contacts and periods. I'd like to be able to represent that two (or more) contacts in different periods are actually the same person (for analytics sake). So if two contacts are linked, they should be treated as one for analytics purposes. For example, I might want to represent that contacts 1 and 3 are the same person.
The difficulty is in representing this concept of "linked" for more than two contacts (e.g. contacts 1, 3, and 4). If there were just two, I could simply have a table with two columns representing the connection. However, since there could be three contacts who are linked (or any number, really), this doesn't work out of the box.
I've come up with two possible solutions for now:
1) Have one table (let's call it linked_contacts) that has two columns, one for each contact ID. This will represent a connection. If there are more than 2 linked contacts, then add a connection to one of the linked contacts. The upside of this is that it's cheap to add another contact into the link. The downside is that in order to get all contacts in the link, I essentially have to construct the graph by making a query for each connection.
2) Still have the linked_contacts table with two columns. When adding a new contact to the link, generate a new row for each connection between the new contact and the pre-existing ones. The upside is that I can get all the linked contacts for a particular contact in a single query. The downside is a larger table.
Additional notes
It's not expected that there will be many links with more than 2 linked contacts, but I would like to support more than 2.
The main use case for retrieving this data is getting all contacts that are linked to a specific contact.
It is very unlikely that a contact will be removed from a link. It will be more common for a contact to be added to a link.
Performance of retrieval is the main factor, since I'll be displaying which contacts are linked to a particular contact on each contact's page.
I'm leaning towards option 2 for ease/speed of retrieval of all contacts in a link given a particular contact. Is the right way to go about it? Are there any other factors that I should take into account? I'm open to other design strategies as well!
I think the "linked list" is a wrong abstraction of your requirements.
If you need to represent the concept of "person" in the database, then just represent it directly, as its own table. For example:
All contacts of the same person share the same PERSON_ID, which implicitly connects them.
Effectively, you turn the CONTACT into a "junction" (aka. "link") table between PERSON and PERIOD, which is a standard way of modeling many-to-many relationship. You can then easily find all persons whose contacts belong to the given period, similarly to this:
SELECT DISTINCT PERSON_ID
FROM CONTACT
WHERE PERIOD_ID = <whatever>
And since there is an index on {PERIOD_ID, PERSON_NO}1, this query can be satisfied by a simple index range scan, which is very fast2.
You can also easily JOIN with the PERSON table if you need to get person's other fields, and/or with PERIOD if you need other fields from there.
1Implicitly created underneath the UNIQUE constraint, indicated by U1 in the diagram above.
2 BTW, you'll also probably need an index in the "opposite" direction: {PERSON_ID, PERIOD_ID}, to satisfy searching for contacts of a given person.
Since you don't expect to have many links, and you are concerned about performance as well as ease of retrieval, and management is insert-heavy, then I would have a separate table with multiple columns representing the links.
CREATE TABLE links (
contact_id0 BIGINT UNSIGNED NULL REFERENCES contacts (id),
contact_id1 BIGINT UNSIGNED NULL REFERENCES contacts (id),
contact_id2 BIGINT UNSIGNED NULL REFERENCES contacts (id),
contact_id3 BIGINT UNSIGNED NULL REFERENCES contacts (id)
);
INSERT INTO links VALUES (1,3,4,NULL);
-- find 3's links
SELECT * WHERE 3 IN (contact_id0,contact_id1,contact_id2,contact_id3);
You might also consider proper normal form of either an edge list or adjacency matrix, but the design above seems to meet your requirements with minimum complexity.
The one design tradeoff is the maximum number of links between contacts is fixed by the number of columns. But as mentioned in my comments, columns are cheap and can be added if demand warrants.
FWIW, if you wanted either your Option 1 or Option 2 and weren't open to other ideas, I would go with your #2.
I'm building a Rails app where a User can have many Addresses. A User also needs a primary billing Address, primary shipping Address, and primary profile Address. These primary fields can point to the same Address. A User cannot have more than one of any primary Address.
I've created a join table called AddressInfo, and I'm bouncing between a few options:
Make 3 columns on the User model corresponding to each of the primary Address ids (this would remove the need for the join model I think).
Add a primary boolean column to AddressInfo, and make sure only one is true when scoped by user_id, address_id and purpose (purpose being billing, shipping or profile).
Add a primary date time column to AddressInfo, and use the most recently updated as the primary address (scoped like option 2).
Maybe these options aren't the best, but it's what I've come up with so far.
Any help on how to resolve this issue would be appreciated.
UPDATE:
To be clear, once an Address is created it should always belong to that User and be undeletable. Ex. a User changes their primary billing address to a new Address: they should still be able to retrieve that old Address (maybe even make it a primary address again). If I go with option 1 and remove the join table, that means I'll need a user_id on Address.
Go with option 1, 3 columns. This will make less of a headache (as a programmer), will run faster, and is more flexible for doing things like combining similar addresses into one. Maybe you have 2 people with the same address, they could share the same record (not recomeneded though).
If you're using rails I would look to a rails solution to 'stay on the rails'.
I would consider STI (Single Table Inheritance)
More info at Rails: Need help defining association for address table
or a Polymorphic relationship - http://guides.rubyonrails.org/association_basics.html#polymorphic-associations
As you want 1 user to have multiple address (rather than ther being multiple user types) than STI may be best for you.
Note: they can also be combined, e.g. http://www.archonsystems.com/devblog/2011/12/20/rails-single-table-inheritance-with-polymorphic-association/
There's a great example of STI addressing this issue at: http://blog.arkency.com/2013/07/sti/
Another possibility is to create another model to hold user.id, addressinfo.id, and an primary_for_address_type (containing "Shipping", "Profile", "Billing").
Constrain this to be unique on user and primary_for_address_type, and you can tag addresses in AddressInfo as being the primary for particular address types in a completely extensible way that still guarantees uniqueness of the primary address.
You might join directly from this model to address, but there's scope for getting it out of sync with your addressinfo model.
i wanna have a Users details stored in the database.. with columns like firstname, last name, username, password, email, cellphone number, activation codes, gender, birthday, occupation, and a few other more. is it good to store all of these on the same table or should i split it between two users and profile ?
If those are attributes of a User (and they are 1-1) then they belong in the user table.
You would only normally split if there were many columns; then you might create another table in a 1-1 mapping.
Another table is obviously required if there are many profile rows per user.
One table should be good enough.
Two tables or more generally vertical portioning comes in when you want to scale out. So you split your tables in multiple tables where usually the partiotioning criteria is the usage i.e., the most common attributes which are used together are housed in one table and others in another table.
One table should be okay. I'd be storing a hash in the password column.
I suggest you read this article on Wikipedia. about database normalization.
It describes the different possibilities and the pros and cons of each. It really depends on what else you want to store and the relationship between the user and its properties.
Ideally one table should be used. If the number of columns becomes harder to manage only then you should move them to another table. In that case, ideally, the two tables should have a one-one relationship which you can easily establish by setting the foreign key in the related table as the primary key:
User
-------------------------------
UserID INT NOT NULL PRIMARY KEY
UserProfile
-------------------------------------------------------
UserID INT NOT NULL PRIMARY KEY REFERENCES User(UserID)
Depend on what kind of application it is, it might be different.
for an enterprise application that my users are the employees as well, I would suggest two tables.
tbl_UserPersonallInformation
(contains the personal information
like name, address, email,...)
tbl_UserSystemInformation (contains
other information like ( Title,
JoinedTheCompanyOn,
LeftTheCompanyOn)
In systems such as "Document Managements" , "Project Information Managements",... this might be necessary.
for example in a company the employees might leave and rejoin after few years and even they will have different job title. The employee had have some activities and records with his old title and he will have some more with the new one. So it should be recorded in the system that with which title (authority) he had done some stuff.