In the previous step, we saved data into items.json. We will move this into a database.
📖 Reference
-
(JA)SQLite入門
We will use a database called SQLite.
- Install SQlite, and make a database file called
mercari.sqlite3. - Open this file and make a table called
items. - Define the items table as follows and save the schema into
db/items.db.- id: int (Identifier unique for each item)
- name: string (Name of the item)
- category: string (Category of the item)
- image_name: string (Filename of the imeage)
Change the endpoints GET /items and POST /items such that items are saved into the database and can be returned based on GET request. Add mercari.sqlite3 to your .gitignore such that it is not committed. items.db should be committed.
🔰 Points
- What are the advantages of saving into a database such as SQLite instead of saving into a single JSON file?
Make an endpoint to return a list of items that include a specified keyword called GET /search.
# Request a list of items containing string "jacket"
$ curl -X GET 'http://127.0.0.1:9000/search?keyword=jacket'
# Expected response for a list of items with name containing "jacket"
{"items": [{"name": "jacket", "category": "fashion"}, ...]}Modify the database as follows. That makes it possible to change the category names without modifying the all categories of items in the items table.
Since GET items should return the category name as before, join these two tables when returning responses.
items table
| id | name | category_id | image_name |
|---|---|---|---|
| 1 | jacket | 1 | 510824dfd4caed183a7a7cc2be80f24a5f5048e15b3b5338556d5bbd3f7bc267.jpg |
| 2 | ... |
categories table
| id | name |
|---|---|
| 1 | fashion |
| ... |
🔰 Points
- What is database normalization?