RunnablePassthrough and RunnableLambda in LangChain

Introduction

RunnablePassthrough and RunnableLambda help you run simple tasks easily without extra setup.

When you want to pass data through without changing it.

When you want to run a small function quickly inside a bigger process.

When you need a simple way to wrap a function to run it later.

When you want to test or debug parts of your code by running small pieces.

When you want to create reusable blocks that do simple jobs.

Syntax

LangChain

from langchain.schema.runnable import RunnablePassthrough, RunnableLambda

# Create a passthrough runnable
passthrough = RunnablePassthrough()

# Create a lambda runnable with a function
lambda_runnable = RunnableLambda(lambda x: x * 2)

# Run them
result1 = passthrough.invoke('hello')
result2 = lambda_runnable.invoke(5)

RunnablePassthrough returns the input as output without changes.

RunnableLambda runs a function you give it on the input.

Examples

This returns 'test' because passthrough sends input as output.

LangChain

passthrough = RunnablePassthrough()
output = passthrough.invoke('test')

This returns 17 because the lambda adds 10 to the input.

LangChain

lambda_runnable = RunnableLambda(lambda x: x + 10)
output = lambda_runnable.invoke(7)

This returns 'HELLO' by converting the input string to uppercase.

LangChain

lambda_runnable = RunnableLambda(lambda s: s.upper())
output = lambda_runnable.invoke('hello')

Sample Program

This program shows how RunnablePassthrough returns the input unchanged, and RunnableLambda runs a function that doubles the input number.

LangChain

from langchain.schema.runnable import RunnablePassthrough, RunnableLambda

# Create a passthrough runnable
passthrough = RunnablePassthrough()

# Create a lambda runnable that doubles numbers
lambda_runnable = RunnableLambda(lambda x: x * 2)

# Use passthrough
result1 = passthrough.invoke('keep this text')
print(f'Passthrough output: {result1}')

# Use lambda runnable
result2 = lambda_runnable.invoke(10)
print(f'Lambda output: {result2}')

Important Notes

RunnablePassthrough is useful when you want to keep data unchanged in a chain.

RunnableLambda lets you quickly wrap any small function to run inside LangChain.

Both help keep your code simple and modular.

Summary

RunnablePassthrough returns input as output without changes.

RunnableLambda runs a small function you provide on the input.

Use them to build simple, reusable pieces in your LangChain workflows.

Practice

1. What does RunnablePassthrough do with the input it receives?

easy

A. Ignores the input and returns a fixed value

B. Transforms the input using a custom function

C. Returns the input exactly as it is without any changes

D. Throws an error if input is not a string

Solution

1. Step 1: Understand RunnablePassthrough behavior

   RunnablePassthrough is designed to return whatever input it receives without modifying it.

2. Step 2: Compare options with behavior

   Only Returns the input exactly as it is without any changes matches this behavior exactly; others describe transformations or errors which RunnablePassthrough does not do.

3. Final Answer:

   Returns the input exactly as it is without any changes -> Option C

4. Quick Check:

   RunnablePassthrough returns input unchanged = A [OK]

Hint: RunnablePassthrough just passes input through unchanged [OK]

Common Mistakes:

• Thinking RunnablePassthrough modifies input
• Confusing RunnablePassthrough with RunnableLambda
• Assuming it throws errors on certain inputs

2. Which of the following is the correct way to create a RunnableLambda that doubles a number input?

easy

A. RunnableLambda(lambda x: x + 2)

B. RunnableLambda(lambda x: x * 2)

C. RunnableLambda(lambda x: x / 2)

D. RunnableLambda(lambda x: x - 2)

Solution

1. Step 1: Identify the doubling function

   Doubling means multiplying the input by 2, so the function should be lambda x: x * 2.

2. Step 2: Match with options

   RunnableLambda(lambda x: x * 2) matches the doubling function exactly; others perform addition, division, or subtraction.

3. Final Answer:

   RunnableLambda(lambda x: x * 2) -> Option B

4. Quick Check:

   Doubling function uses multiplication by 2 = D [OK]

Hint: Doubling means multiply input by 2 in lambda [OK]

Common Mistakes:

• Using addition instead of multiplication
• Confusing division or subtraction for doubling
• Incorrect lambda syntax

3. What will be the output of this code?

passthrough = RunnablePassthrough()
lambda_runner = RunnableLambda(lambda x: x.upper())
result = lambda_runner.invoke(passthrough.invoke('hello'))
print(result)

medium

A. 'HELLO'

B. 'hello'

C. Error: RunnablePassthrough cannot be invoked

D. 'Hello'

Solution

1. Step 1: Trace RunnablePassthrough output

   Calling passthrough.invoke('hello') returns 'hello' unchanged.

2. Step 2: Apply RunnableLambda function

   The lambda converts input to uppercase, so 'hello'.upper() returns 'HELLO'.

3. Final Answer:

   'HELLO' -> Option A

4. Quick Check:

   Passthrough returns input, lambda uppercases it = C [OK]

Hint: Passthrough returns input, lambda transforms it [OK]

Common Mistakes:

• Expecting passthrough to modify input
• Confusing case conversion result
• Assuming runtime error on invoke

4. Identify the error in this code snippet:

lambda_runner = RunnableLambda(lambda x: x + 1)
result = lambda_runner.invoke('5')
print(result)

medium

A. TypeError because string '5' cannot be added to integer 1

B. SyntaxError in lambda function

C. RunnableLambda cannot be invoked with strings

D. No error; output will be '51'

Solution

1. Step 1: Analyze lambda operation on input

   The lambda tries to add 1 to input '5', which is a string, causing a type mismatch.

2. Step 2: Identify error type

   Adding integer 1 to string '5' raises a TypeError in Python.

3. Final Answer:

   TypeError because string '5' cannot be added to integer 1 -> Option A

4. Quick Check:

   Adding int to string causes TypeError = B [OK]

Hint: Adding int to string causes TypeError in Python [OK]

Common Mistakes:

• Assuming implicit string to int conversion
• Thinking output is string concatenation
• Confusing error types

5. You want to create a LangChain workflow that takes a list of numbers, passes it unchanged, then doubles each number. Which combination of RunnablePassthrough and RunnableLambda is correct?

hard

A. Use RunnablePassthrough twice, no lambda needed

B. Use RunnableLambda with lambda x: x, then RunnablePassthrough to double numbers

C. Use RunnableLambda with lambda x: x*2, then RunnablePassthrough to pass list

D. Use RunnablePassthrough to pass the list, then RunnableLambda with lambda x: [i*2 for i in x]

Solution

1. Step 1: Pass list unchanged with RunnablePassthrough

   RunnablePassthrough returns the list as is, so it fits the first step.

2. Step 2: Double each number with RunnableLambda

   RunnableLambda with lambda x: [i*2 for i in x] correctly doubles each element in the list.

3. Final Answer:

   Use RunnablePassthrough to pass the list, then RunnableLambda with lambda x: [i*2 for i in x] -> Option D

4. Quick Check:

   Passthrough passes list, lambda doubles elements = A [OK]

Hint: Passthrough passes input; lambda transforms list elements [OK]

Common Mistakes:

• Trying to double list with passthrough
• Using lambda that multiplies list object, not elements
• Reversing order of passthrough and lambda