When to use pipelines in Redis - Step-by-Step Execution

Concept Flow - When to use pipelines

Start: Need to send multiple commands

↓

Send commands one-by-one

→High network latency, slow

↓

Use pipeline to batch commands

↓

Send all commands at once

↓

Receive all responses together

↓

Process responses

↓

End

When you have many commands to send to Redis, using a pipeline batches them to reduce network delays and speed up execution.

Execution Sample

Redis

pipeline = redis_client.pipeline()
pipeline.set('key1', 'value1')
pipeline.get('key1')
pipeline.set('key2', 'value2')
results = pipeline.execute()

This code batches three commands into a pipeline and sends them together to Redis, then collects all results at once.

Execution Table

Step	Action	Command Sent	Network Calls	Response Received	Notes
1	Add SET key1	SET key1 value1	0 (queued)	No	Command queued in pipeline
2	Add GET key1	GET key1	0 (queued)	No	Command queued in pipeline
3	Add SET key2	SET key2 value2	0 (queued)	No	Command queued in pipeline
4	Execute pipeline	All commands sent together	1 network call	Yes, all responses	Batch sent, responses received together
5	Process results	N/A	N/A	['OK', 'value1', 'OK']	Results correspond to commands in order
6	End	N/A	N/A	N/A	Pipeline execution complete

💡 All commands sent in one network call, reducing latency compared to sending individually

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	Final
pipeline.commands	[]	[SET key1 value1]	[SET key1 value1, GET key1]	[SET key1 value1, GET key1, SET key2 value2]	[sent all commands]	[commands cleared after execute]
results	None	None	None	None	['OK', 'value1', 'OK']	['OK', 'value1', 'OK']

Key Moments - 3 Insights

Why don't we see network calls when adding commands to the pipeline?

Commands are queued locally in the pipeline object (see steps 1-3 in execution_table), so no network calls happen until execute() is called.

How do we know the responses match the commands?

Responses come back in the same order as commands were queued (step 5), so each result corresponds to its command.

What happens if we send commands one-by-one instead of using a pipeline?

Each command causes a separate network call, increasing latency and slowing performance (see step 1 in concept_flow).

Visual Quiz - 3 Questions

Look at the execution_table, at which step are all commands sent to Redis in one network call?

A Step 4

B Step 2

C Step 1

D Step 5

💡 Hint

Check the 'Network Calls' column to see when commands are sent

According to variable_tracker, what is the state of 'results' after step 3?

A ['OK', 'value1', 'OK']

B ['value1']

C None

D Empty list

💡 Hint

Look at the 'results' row under 'After Step 3' in variable_tracker

If we did not use a pipeline, how would the number of network calls change compared to step 4?

A It would be 1 call

B It would be 3 calls

C It would be 0 calls

D It would be 2 calls

💡 Hint

Refer to concept_flow where sending commands one-by-one causes multiple network calls

Concept Snapshot

Use Redis pipelines to batch multiple commands.
Queue commands locally, then send all at once.
Reduces network latency by minimizing round trips.
Responses come back in the same order.
Ideal when sending many commands quickly.

Full Transcript

When you have many Redis commands to send, sending each one separately causes many network calls, which slows things down. Instead, you can use a pipeline to queue commands locally. When you call execute on the pipeline, all commands are sent together in one network call. Redis processes them and sends back all responses at once, in the same order. This reduces network delays and speeds up your program. The example code shows adding three commands to a pipeline and then executing them together. The execution table traces each step: commands are queued first without network calls, then all sent at once, and finally results are received and processed. The variable tracker shows how the pipeline commands list grows and then clears after execution, and how results appear only after sending. Key moments clarify why commands are queued before sending, how responses match commands, and why sending commands one-by-one is slower. The quiz tests understanding of when commands are sent, the state of results before execution, and the difference in network calls with and without pipelines.