1. Updating Data (The PATCH method)#

While PUT is used to replace a resource entirely, PATCH is used for partial updates. This is more efficient because it reduces the amount of data sent over the network.

@app.patch("/items/{item_id}", response_model=ItemResponse)
def update_item(item_id: int, item_update: ItemUpdate, db: Session = Depends(get_db)):
    # 1. Find the item
    db_item = db.query(models.ItemDB).filter(models.ItemDB.id == item_id).first()
    if not db_item:
        raise HTTPException(status_code=404, detail="Item not found")
    
    # 2. Update only the fields provided
    update_data = item_update.dict(exclude_unset=True)
    for key, value in update_data.items():
        setattr(db_item, key, value)
    
    # 3. Save and return
    db.commit()
    db.refresh(db_item)
    return {"message": "Item updated", "item": db_item}

2. Deleting Data#

Deleting a record is straightforward, but it’s critical to verify the record exists first to avoid database errors.

@app.delete("/items/{item_id}")
def delete_item(item_id: int, db: Session = Depends(get_db)):
    db_item = db.query(models.ItemDB).filter(models.ItemDB.id == item_id).first()
    if not db_item:
        raise HTTPException(status_code=404, detail="Item not found")
    
    db.delete(db_item)
    db.commit()
    return {"message": "Item deleted successfully"}

3. Soft Delete vs. Hard Delete#

Today I learned that in production, we often use Soft Deletes. Instead of removing the row, we add a is_deleted column. This prevents accidental data loss and keeps a history for audits. For this challenge, we are using Hard Deletes (removing the row) to keep things simple.

🛠️ Implementation Checklist#

• Implemented a PATCH endpoint for partial updates.
• Implemented a DELETE endpoint.
• Used exclude_unset=True in Pydantic to handle partial updates correctly.
• Verified that deleted items no longer appear in the GET /items/ list.

📚 Resources#

1. Official Docs: FastAPI Update Data
2. Book: FastAPI: Modern Python Web Development (Chapter 6: Finalizing CRUD).

With this, Phase 3 is complete!