Using Redis commands

Learn how to interact with Redis using commands.

Client applications and tools interact with Redis using commands. Most of the commands implement data types to store and retrieve data, but there are also commands that deal with server configuration, security, and more.

The sections below give an overview of how Redis commands operate. See the Redis commands reference for a complete list of commands.

Command structure

Each command is identified by a unique name. Related groups of commands tend to follow a consistent naming convention. For example, all commands that deal with hashes start with the H prefix. Most commands receive one or more arguments that specify the data to operate on. For data type commands, the first argument is usually the key that identifies the target data object.

After you issue a command, the server attempts to process it and then returns a response. Commands that update data typically return a status message (such as OK) or a number that indicates the number of items changed or updated. Commands that retrieve data return the requested data. An unsuccessful command returns an error message that describes the problem.

Interacting with a Redis server involves a sequence of commands and responses. The effect of a given command is the same regardless of whether you send it from a client library, or from a client tool such as redis-cli or Redis Insight. This is very useful during development. You can use a high-level tool to experiment with a command, set up test data, or prototype a data model, and then access the prepared data from your application code. Most Redis code examples are presented with an excerpt of a CLI session and the equivalent application code for each client library.

Batching commands

Although you can issue Redis commands one at a time, it's often more efficient to batch a sequence of related commands together into a pipeline. A pipeline sends several commands to the server as a single communication and receives the responses in the same way. See Pipelining for a full description of the technique and see also the pipelining examples for the client libraries.

Another reason to batch commands is to treat them as an uninterrupted unit. You should do this if you need to be sure that the commands are all completed without the same data being modified by another client (which could leave the data in an inconsistent state). Redis uses transactions to implement this behavior. See Transactions for more information and see also the transaction examples for the client libraries.

More information

The other pages in this section describe Redis command concepts in more detail: