How to keep sensitive fields when processing Redis stored JSON strings in Go?

Go language handles JSON strings stored in Redis: protect sensitive fields

In Go projects, special attention should be paid to data security when processing JSON data containing sensitive information (such as passwords) and storing it in Redis. Although using json:"-" tag directly prevents sensitive fields from being exposed in response to clients, it will lose the necessary information stored in Redis. This article discusses several solutions to deal with this problem and analyzes its advantages and disadvantages.

Solution 1: Use two structures

A common solution is to define two structures: one for client response and the other for internal data storage. The client response structure uses json:"-" to hide sensitive fields; the internal storage structure contains all fields. This method is clear and easy to understand, but requires maintenance of two structures and the code is redundant.

Solution 2: Customize JSON tags and serialization functions

To reduce code redundancy, you can customize JSON tags (for example, json:"hideWhenNetworkResponse" ) and write two functions:

• MarshalForClient(data interface{}) ([]byte, error) : Serialized data is used for client responses, filtering sensitive fields based on custom tags.
• MarshalForRedis(data interface{}) ([]byte, error) : Serialized data is used for Redis storage, retaining all fields.

Example structure:

 type Data struct {
    Name string
    Password string `json:"hideWhenNetworkResponse"`
}

This approach is more elegant, but requires careful management of custom tags to avoid missed or incorrect use.

Scheme 3: Prioritize the use of explicit codes that handle sensitive fields separately

Although the custom tag method looks more concise, the author recommends writing clear code directly to handle sensitive fields. This approach is easier to understand and maintain, reducing the risk of security issues due to improper use of labels. When serialized data is used for client response, sensitive fields are explicitly removed or replaced; when stored in Redis, all fields are preserved.

Which option you choose ultimately depends on the complexity of the project and the team's preferences. For small projects, the clarity of solution 3 may be more advantageous; for large projects, the code reusability of solution 2 may be more efficient. The key is to choose a way to process sensitive data that is easy to understand, maintain and secure.

The above is the detailed content of How to keep sensitive fields when processing Redis stored JSON strings in Go?. For more information, please follow other related articles on the PHP Chinese website!