ReviseAlgo Logo

Iterators & Generators

yield from

Delegating to sub-generators with yield from — simplifying nested iteration, bidirectional communication, and building recursive generators.

Interview: Advanced generator pattern — shows deep understanding of Python generator mechanics.

Last Updated: June 12, 2026 • 7 min read

yield from (Python 3.3+) delegates to a sub-generator, transparently forwarding all values, sends, and exceptions. It replaces the boilerplate of iterating over a sub-generator and yielding each value individually.

Basic Usage

  • yield from iterable — yields all items from the iterable one by one
  • Works with any iterable, not just generators — lists, ranges, strings, etc.
  • Replaces the pattern: for item in sub: yield item

Bidirectional Communication

  • yield from transparently forwards send() and throw() to the sub-generator
  • The return value of the sub-generator becomes the result of the yield from expression
  • This enables coroutines to compose and communicate in complex ways

Recursive Generators

yield from is essential for recursive generators, especially tree traversal:

  • Tree traversal — yield from left subtree, yield current, yield from right subtree
  • Flattening nested structures — recursively yield from sub-lists
  • Without yield from, recursive generators need explicit for loops

Interview Insight

Know when to use yield from vs a for loop. yield from is cleaner and properly forwards send/throw/close to the sub-generator — important for coroutine composition.

Use Cases

Tree traversal — in-order, pre-order, post-order traversal of binary trees

Flattening nested structures — lists of lists, nested dicts

Coroutine composition — chaining generators with bidirectional communication

Parser combinators — delegating to sub-parsers

Graph algorithms — DFS/BFS traversal with generator-based exploration

Common Mistakes

Forgetting that yield from captures the sub-generator return value

Using yield from with a non-iterable — raises TypeError

Not understanding that send/throw/close are forwarded to the sub-generator

Using for loop + yield instead of yield from — works but misses send/throw forwarding

Infinite recursion in recursive generators — must have a base case