12 KiB
HELLO - Start Building Your Own ORM
Use PHP to reshape the kinds of objects coming from a SQL database
Learning Objectives
- Learn about the challenges of converting SQL rows into JSON
- Practice using PHP to reshape objects
- etc.
Prerequisites
- Experience with RESTful routes
- Basic PHP
- Familiarity with JSON
- Familiarity with SQL
Getting Started
- You'll be working in one PHP file, and building out your answers as you work through this hw.
Deliverables
Technical Requirements
- Must run without syntax errors (comment out code that doesn't work and add comments)
- Must complete parts 1 - 6
Submission Guidelines
- Must be submitted no later than before the start of next class
Context
The founder of the start-up H.E.L.L.O (Housing Elite Leveraged Location Organization) wants you to help launch her amazing idea.
Too many college graduates these days have too high loans to pay back. This is preventing them from being able to get into premium housing locations, where everything is in walking distance and work is a sweet sweet short commute away.
One night, our brilliant entrepreneur was reminiscing about how much she enjoyed living in a dorm and thought, why can't people keep on living in a dorm?
And thus H.E.L.L.O. was founded. A premier housing solution. For the low low starting price of $3999 a month for a bunk bed shared with one other person and a co-ed bathroom shared by everyone on the same floor, anyone's housing dreams can come true!
H.E.L.L.O already has funding for one location!
Your first task is to help build a tenant tracker.
Part 1 - explanation
SQL stores data in rows and columns, so in order to get the data to display on a web page, we need to convert the data from rows and columns into associative arrays with keys and values. Fortunately, PHP does this for us, but we need to do some manual coding to get it to work exactly as we want it to.
If we just have one table with the following keys and values:
| name | age | home_id |
|---|---|---|
| Chase | 30 | 1 |
PHP can turn this into an associative array which would look something like this:
[
[
"name" => "Chase",
"age" => "30",
"home_id" => "1"
]
]
Remember, that PHP doesn't allow us to spontaneously create an object with any arbitrary properties that we want like JavaScript does. We can use associative in this manner, though. It's a good way to keep track of a data structure that requires key/value pairs.
If we had more than one person...
| name | age | home_id |
|---|---|---|
| Chase | 30 | 1 |
| Gert | 23 | 1 |
PHP will give us something like this:
[
[
"name" => "Chase",
"age" => "30",
"home_id" => "1"
],
[
"name" => "Gert",
"age" => "23",
"home_id" => "1"
]
]
This is straightforward. However, we start to get a few wrinkles once we start joining tables. In this homework's situation, our location HAS MANY people (and each person HAS ONE location). This is also called a ONE to MANY relationship. If we were to join Chase's info with his location's info, our table could look like the following:
| name | age | home_id | home_street | home_city | home_state | home_zip_code |
|---|---|---|---|---|---|---|
| Chase | 30 | 1 | "1600 Broadway" | "New York" | "NY" | 10019 |
PHP would turn this all this info into just one associative array:
Original Format
[
[
"name"=>"Chase",
"age"=>"30",
"home_id"=>"1",
"home_street"=>"1600 Broadway",
"home_city"=>"New York",
"home_state"=>"NY",
"home_zip_code"=>"10019"
]
]
But what we really want is to change the shape of the data and have home as a nested associative array. We'd also want to convert our numbers back into numbers from strings:
New Format
[
"name"=>"Chase",
"age"=>30,
"home"=> [
"id"=> 1,
"street"=>"1600 Broadway",
"city"=>"New York",
"state"=>"NY",
"zip_code"=> 10019
]
]
This might seem relatively minor, but it makes more sense to store the data this way. It's much more representative of the real world.
Part 2 - code along
First, let's define a new method data_transform. It will take a single hash inside of an array called person_array:
def data_transform(person_array)
end
Now let's create a fake people array, containing a single hash. This will represent what the PG gem will give us from Postgres when we use it in rails. Next, call data_transform and pass that fake people array in as a parameter:
people = [
{
"name"=>"Chase",
"age"=>"30",
"home_id"=>"1",
"home_street"=>"1600 Broadway",
"home_city"=>"New York",
"home_state"=>"NY",
"home_zip_code"=>"10019"
}
]
def data_transform(person_array)
end
p data_transform people
Second, inside of data_transform, let's make a new object called person
def data_transform(person_array)
person = {}
end
We can make whatever keys we want for our new object:
def data_transform(person_array)
person = {
"name" => "Chase",
"age" => "30"
}
end
And we can ALMOST place our values from our person_array into our new hash
person = {
"name" => person_array["age"],
"age" => person_array["age"]
}
Check it! This won't quite work because our person_array is an array. We can easily take the single hash out of the array using a ruby method .first
def data_transform(person_array)
person = person_array.first
return {
"name" => person["name"],
"age" => person["age"]
}
end
Now try again! The full code:
people = [
{
"name"=>"Chase",
"age"=>"30",
"home_id"=>"1",
"home_street"=>"1600 Broadway",
"home_city"=>"New York",
"home_state"=>"NY",
"home_zip_code"=>"10019"
}
]
def data_transform(person_array)
person = person_array.first
return {
"name" => person["name"],
"age" => person["age"]
}
end
p data_transform people
should print a hash containing just the name and the age of the person:
{"name"=>"Chase", "age"=>"30"}
It ALMOST matches the New Format, but the age is a string. We can use a method to fix that in a snap!
return {
"name" => person["name"],
"age" => person["age"].to_i
}
Check it again!
{"name"=>"Chase", "age"=>30}
Part 3 - build it
Don't forget, we want our data_transform function to return an object that also includes the other information about our person.
{
"name"=>"Chase",
"age"=>30,
"home"=> {
"id"=> 1,
"street"=>"1600 Broadway",
"city"=>"New York",
"state"=>"NY",
"zip_code"=> 10019
}
}
How do we get that nested home hash in the results?
Part 4 - build it
As we get new clients, not all of them will have a home right away. We need to add a way to check if a person has a home address, and if true, return the reshaped hash else just return the person hash
Don't forget to test it, there is sample 'data' available in Part 1 for you to use
Part 5 - build it
HELLO already has three tenants! Converting each object's format one by one is a drag. Write a new function data_transforms that converts an array of many people hashes into the New Format shape:
Original Format
people = [
{
"name"=>"Gert",
"age"=>"23",
"home_id"=> "1",
"home_street"=>"1600 Broadway",
"home_city"=>"New York",
"home_state"=>"NY",
"home_zip_code"=> "10019"
},
{
"name"=>"Alex",
"age"=>"42",
"home_id"=> "1",
"home_street"=>"1600 Broadway",
"home_city"=>"New York",
"home_state"=>"NY",
"home_zip_code"=> "10019"
},
{
"name"=>"Nico",
"age"=>"61",
"home_id"=> "1",
"home_street"=>"1600 Broadway",
"home_city"=>"New York",
"home_state"=>"NY",
"home_zip_code"=> "10019"
}
]
Check yourself: does your method still work if a person doesn't yet have a home address?
New Format
people = [
{
"name"=>"Gert",
"age"=>23,
"home"=> {
"id"=> 1,
"street"=>"1600 Broadway",
"city"=>"New York",
"state"=>"NY",
"zip_code"=> 10019
}
},
{
"name"=>"Alex",
"age"=>42,
"home"=> {
"id"=> 1,
"street"=>"1600 Broadway",
"city"=>"New York",
"state"=>"NY",
"zip_code"=> 10019
}
},
{
"name"=>"Nico",
"age"=> 61,
"home"=> {
"id"=> 1,
"street"=>"1600 Broadway",
"city"=>"New York",
"state"=>"NY",
"zip_code"=> 10019
}
}
]
Part 6
For this part, write a brand new data_transforms_locations method. Save your previous one(s)
HELLO is growing rapidly and has acquired an amazing high profile space for their next luxury dorm experience. So next, we'd like to be able to show a list of all our addresses with their corresponding tenants:
Original Format
NOTE assume all people with home_id of 1 are next to each other in the array. Same with home_id of 2, 3, etc
people = [
{
"name"=>"Gert",
"age"=>"23",
"home_id"=> "1",
"home_street"=>"1600 Broadway",
"home_city"=>"New York",
"home_state"=>"NY",
"home_zip_code"=> "10019"
},
{
"name"=>"Alex",
"age"=>"42",
"home_id"=> "1",
"home_street"=>"1600 Broadway",
"home_city"=>"New York",
"home_state"=>"NY",
"home_zip_code"=> "10019"
},
{
"name"=>"Nico",
"age"=>"61",
"home_id"=> "1",
"home_street"=>"1600 Broadway",
"home_city"=>"New York",
"home_state"=>"NY",
"home_zip_code"=> "10019"
},
{
"name"=>"Molly",
"age"=>"55",
"home_id"=> "2",
"home_street"=>"1 Alcatraz Island",
"home_city"=>"San Francisco",
"home_state"=>"CA",
"home_zip_code"=> "94133"
},
{
"name"=>"Karolina",
"age"=>"82",
"home_id"=> "2",
"home_street"=>"1 Alcatraz Island",
"home_city"=>"San Francisco",
"home_state"=>"CA",
"home_zip_code"=> "94133"
}
]
New Format
[
{
"id"=> 1,
"street"=>"1600 Broadway",
"city"=>"New York",
"state"=>"NY",
"zip_code"=> 10019,
"tenants" => [
{
"name"=>"Gert",
"age"=>23
},
{
"name"=>"Alex",
"age"=>42
},
{
"name"=>"Nico",
"age"=>61
}
]
},
{
"id"=> 2,
"street"=>"1 Alcatraz Island",
"city"=>"San Francisco",
"state"=>"CA",
"zip_code"=> 94133,
"tenants" => [
{
"name"=>"Molly",
"age"=>23
},
{
"name"=>"Karolina",
"age"=>42
}
]
}
]
Part 7 - Hungry for more
Create the ability for data_transform_location to optionally return a specific home (with its tenants), based on an optional second home_id param:
p data_transform_location people, 2
should return
{
"id"=> 2,
"street"=>"1 Alcatraz Island",
"city"=>"San Francisco",
"state"=>"CA",
"zip_code"=> 94133,
"tenants" => [
{
"name"=>"Molly",
"age"=>23
},
{
"name"=>"Karolina",
"age"=>42
}
]
}
You did it! You started building your own ORM (Object Relational Mapping)
Copyright 2019, General Assembly Space. Licensed under CC-BY-NC-SA, 4.0