Data Loading Node

What is a Data Loading Node?

The Data Loading Node is a tool used to read previously stored data from a persistent database. It acts like a smart retriever that can quickly find and extract information previously saved in a persistent database based on the "data key" you provide, for use by other nodes in the workflow.

Image Description:

The Data Loading Node interface primarily consists of a query condition area, including scope selection, data key input field, and other key elements. Users can retrieve previously stored data from the persistent database by configuring these parameters.

Why do you need a Data Loading Node?

In intelligent workflows, there is often a need to transfer and use data across different times or different sessions. For example:

• Remember user preferences and use them directly in the next conversation
• Store key information from the last interaction for subsequent processing
• Save business data, allowing retrieval and use at any time in the future

The Data Loading Node is designed to address this need, allowing you to conveniently retrieve any previously stored information, establishing a "long-term memory" capability for the workflow, giving the AI assistant persistent data access capabilities.

Application Scenarios

The Data Loading Node is suitable for the following scenarios:

1. User Setting Memory: Read user preferences, usage habits, and other information to provide personalized services
2. Business Process Continuity: Maintain business continuity across multiple conversations, such as reading information about previously unfinished orders
3. Knowledge Retrieval: Read previously saved specialized knowledge or rules from persistent storage
4. User Identity Recognition: Read user identity information for subsequent permission control and personalized services

Node Parameter Description

Input Parameters

Parameter Name	Parameter Description	Required	Default Value	Notes
Scope	The storage range of the data	Yes	Current Topic	Determines which range to look for data in, available ranges include: Current Topic, Global, etc.
Data Key	The identifier of the data to be read	Yes	None	A unique identifier used to find data, supports using variables, such as "@UserID"

Output Parameters

After successful execution, the Data Loading Node will output the following variables, which can be used in subsequent nodes:

Output Variable Name	Data Type	Description
Value (value)	String/Array	Depending on the data type at storage time, this could be a simple text string or a complex JSON object or array

Usage Instructions

Basic Configuration Steps

1. Select Scope: Choose the data scope to query from the dropdown menu, usually with options like "Current Topic", "Current User", etc.
2. Set Data Key: Enter the identifier of the data to be read in the "Data Key" input box
   1. You can directly input text, such as "UserPreferences"
   2. You can also click the "@" button to select a variable from the variable list as the data key
3. Connect Subsequent Nodes: Connect the output of the Data Loading Node to subsequent nodes that need to use the data

Advanced Techniques

1. Dynamic Data Keys: When you need to read different data based on different situations, you can use variables as data keys. For example, you can use "@UserID" as a data key, and the system will automatically read the corresponding data based on the current user ID.
2. Combined with Conditional Judgment: After loading data, you can use conditional judgment nodes to check whether the data exists and is valid, thereby building more complex logical flows.
3. Data Transformation: If the data read is in JSON format, you can use code nodes for parsing and conversion to extract specific fields.

Precautions

Handling Non-existent Data

When the data key to be read does not exist in the database, the Data Loading Node will output an empty value. In subsequent nodes, it is recommended to first check whether the value is empty to avoid process errors due to non-existent data.

Data Expiration Issues

If an expiration time was set when the data was stored, the data will automatically become invalid after that time. Please ensure appropriate alternative processing plans are in place in case the data may expire.

Data Type Consistency

The Data Loading Node will return data of the same type as when it was stored. For example, if a JSON object was stored, a JSON object will also be obtained when loading. Please ensure that subsequent nodes can correctly process that type of data.

Frequently Asked Questions

Question 1: Why can't I read the previously stored data?

Possible reasons:

• Data key name is incorrect or inconsistently spelled
• Wrong scope selection (e.g., data stored in "Global" scope but trying to read from "Current Topic")
• Data has expired (if an expiration time was set during storage)
• Data may have been deleted or overwritten by other processes

Solutions:

• Confirm that the data key name is exactly the same as when it was stored
• Check if the scope is the same as when it was stored
• If you suspect the data has expired, try storing it again first
• Add log or debug nodes after the Data Loading Node to check the loading results

Question 2: How to determine if data was successfully loaded?

Solution:

Add a conditional judgment node after the Data Loading Node to check if the output value is empty. If it's not empty, then the data was successfully loaded; if it's empty, then the data may not exist or has expired.

Common Node Combinations

Node Type	Combination Reason
Conditional Judgment Node	Make judgments after loading data to decide subsequent process
Code Execution Node	Process and transform complex data after reading
Large Model Call Node	Pass loaded data as context to the large model, enhancing the relevance of model answers
Data Storage Node	First use the Data Storage Node to save data, later use the Data Loading Node to read it