Pagination (PageNumber, Cursor, Limit/Offset) in Django - Performance & Optimization

Performance: Pagination (PageNumber, Cursor, Limit/Offset)

MEDIUM IMPACT

Pagination affects how much data is loaded and rendered at once, impacting page load speed and responsiveness.

Loading large lists of items in a web page

Django

def get_items(request):
    cursor = request.GET.get('cursor')
    limit = 100
    if cursor:
        items = Item.objects.filter(id__gt=cursor).order_by('id')[:limit]
    else:
        items = Item.objects.all().order_by('id')[:limit]
    return render(request, 'items.html', {'items': items})

Cursor pagination uses indexed queries for faster, stable data retrieval and reduces load time.

📈 Performance GainReduces query time by 50%+; lowers reflows by loading fewer items per page

Loading large lists of items in a web page

Django

def get_items(request):
    page = int(request.GET.get('page', 1))
    limit = 100
    offset = (page - 1) * limit
    items = Item.objects.all()[offset:offset+limit]
    return render(request, 'items.html', {'items': items})

Using limit/offset pagination on large datasets causes slow database queries and inconsistent results when data changes.

📉 Performance CostBlocks rendering for 200+ ms on large pages; triggers multiple reflows if many items render

Performance Comparison

Pattern	DOM Operations	Reflows	Paint Cost	Verdict
Limit/Offset Pagination on large data	High (many nodes)	Multiple reflows per page	High paint cost	[X] Bad
Cursor Pagination with indexed queries	Low (limited nodes)	Single reflow per page	Low paint cost	[OK] Good

Rendering Pipeline

Pagination controls how much data is fetched and rendered, affecting layout and paint stages by limiting DOM nodes.

→Layout

→Paint

→Composite

⚠️ BottleneckLayout stage due to many DOM nodes when loading large pages

Core Web Vital Affected

LCP

Pagination affects how much data is loaded and rendered at once, impacting page load speed and responsiveness.

Optimization Tips

1Avoid large offset values in limit/offset pagination to reduce slow queries.

2Use cursor pagination with indexed columns for faster, stable data retrieval.

3Limit the number of items rendered per page to reduce layout and paint costs.

Performance Quiz - 3 Questions

Which pagination method generally improves page load speed for large datasets?

A Loading all data at once without pagination

B Cursor pagination using indexed queries

C Limit/offset pagination with large offsets

D Using random page numbers without order

DevTools: Performance

How to check: Record a page load with pagination, observe the flame chart for scripting and rendering times, and check the number of layout and paint events.

What to look for: Look for long scripting or layout times and many reflows indicating inefficient pagination.

Practice

1. Which Django REST Framework pagination style uses a page number to fetch specific pages of data?

easy

A. PageNumberPagination

B. CursorPagination

C. LimitOffsetPagination

D. OffsetPagination

Solution

1. Step 1: Understand pagination styles

   PageNumberPagination uses page numbers like 1, 2, 3 to get data pages.

2. Step 2: Match style to description

   CursorPagination uses a cursor token, LimitOffsetPagination uses limit and offset numbers, so they don't use page numbers.

3. Final Answer:

   PageNumberPagination -> Option A

4. Quick Check:

   Page number style = PageNumberPagination [OK]

Hint: PageNumberPagination uses simple page numbers like 1, 2, 3 [OK]

Common Mistakes:

• Confusing CursorPagination with page numbers
• Thinking LimitOffsetPagination uses page numbers
• Assuming OffsetPagination is a valid DRF style

2. Which of the following is the correct way to set LimitOffsetPagination in Django REST Framework settings?

easy

A. "DEFAULT_PAGINATION_CLASS": "rest_framework.pagination.OffsetPagination"

B. "DEFAULT_PAGINATION_CLASS": "rest_framework.pagination.PageNumberPagination"

C. "DEFAULT_PAGINATION_CLASS": "rest_framework.pagination.LimitOffsetPagination"

D. "DEFAULT_PAGINATION_CLASS": "rest_framework.pagination.CursorPagination"

Solution

1. Step 1: Identify correct class path

   LimitOffsetPagination is located at rest_framework.pagination.LimitOffsetPagination.

2. Step 2: Verify syntax for settings

   The setting key is DEFAULT_PAGINATION_CLASS and the value is the full class path as a string.

3. Final Answer:

   "DEFAULT_PAGINATION_CLASS": "rest_framework.pagination.LimitOffsetPagination" -> Option C

4. Quick Check:

   Correct class path and setting key = "DEFAULT_PAGINATION_CLASS": "rest_framework.pagination.LimitOffsetPagination" [OK]

Hint: Use full class path string for pagination in settings [OK]

Common Mistakes:

• Using OffsetPagination which does not exist
• Missing quotes around class path string
• Mixing pagination class names

3. Given this code snippet using CursorPagination, what will be the value of the next cursor if the current cursor is 'abc123' and page size is 2?

class MyCursorPagination(CursorPagination):
    page_size = 2

paginator = MyCursorPagination()
next_cursor = paginator.get_next_link()

medium

A. A tuple with offset and limit values

B. An integer representing the next page number

C. None, because get_next_link() returns nothing

D. A URL containing a cursor parameter with a new encoded cursor

Solution

1. Step 1: Understand CursorPagination behavior

   CursorPagination returns URLs with encoded cursor tokens for next pages, not page numbers or tuples.

2. Step 2: Analyze get_next_link() output

   get_next_link() returns a URL string containing the next cursor parameter for pagination.

3. Final Answer:

   A URL containing a cursor parameter with a new encoded cursor -> Option D

4. Quick Check:

   CursorPagination next link = URL with cursor [OK]

Hint: CursorPagination returns URLs with cursor tokens, not numbers [OK]

Common Mistakes:

• Expecting page numbers from CursorPagination
• Thinking get_next_link() returns None
• Confusing limit/offset with cursor

4. You have this Django REST Framework view using LimitOffsetPagination but it raises an error:

class MyLimitOffsetPagination(LimitOffsetPagination):
    default_limit = '10'

class MyView(ListAPIView):
    pagination_class = MyLimitOffsetPagination
    queryset = MyModel.objects.all()
    serializer_class = MySerializer

What is the likely cause of the error?

medium

A. pagination_class should be a string path, not a class

B. default_limit should be an integer, not a string

C. ListAPIView does not support pagination

D. queryset must be a list, not a QuerySet

Solution

1. Step 1: Check default_limit type

   default_limit must be an integer, but it is set as a string '10', causing a type error.

2. Step 2: Verify other parts

   pagination_class can be a class, ListAPIView supports pagination, queryset can be a QuerySet.

3. Final Answer:

   default_limit should be an integer, not a string -> Option B

4. Quick Check:

   default_limit type error = default_limit should be an integer, not a string [OK]

Hint: default_limit must be int, not quoted string [OK]

Common Mistakes:

• Setting default_limit as string instead of int
• Thinking pagination_class must be string path
• Assuming ListAPIView disables pagination

5. You want to implement pagination for a large dataset where new items are frequently added. Which pagination style is best to avoid duplicate or missing items when users navigate pages?

hard

A. CursorPagination

B. PageNumberPagination

C. LimitOffsetPagination

D. No pagination

Solution

1. Step 1: Understand pagination challenges with dynamic data

   PageNumber and LimitOffset can cause duplicates or missing items if data changes between requests.

2. Step 2: Identify pagination style that handles dynamic data well

   CursorPagination uses a stable cursor based on item order, preventing duplicates or skips when data changes.

3. Final Answer:

   CursorPagination -> Option A

4. Quick Check:

   Dynamic data needs CursorPagination [OK]

Hint: Use CursorPagination for changing data to avoid duplicates [OK]

Common Mistakes:

• Choosing PageNumberPagination for dynamic data
• Thinking LimitOffsetPagination handles data changes well
• Ignoring pagination for large datasets