SQL Explore, Query, Visualize in Chat
Overview
Athena can answer questions using data autonomously.
To start an SQL workflow in Athena, select specific datasets and enable the SQL Explore, Query, Visualize tool within the chat interface.
Providing precise context in your initial request and follow-up questions can greatly assist Athena in understanding and processing your request accurately.
Set up datasets and tools
Connect Datasets
Before starting SQL workflows in Chat, the first step is to connect your datasets. This ensures that Athena has the data it needs to perform tasks.
To select datasets for your SQL workflows, open the dataset section to the right of the chat input. Here, you can choose the datasets you want to work with.
Select Up to Five Datasets
You can select up to five datasets for each SQL workflow to manage complexity and maintain performance.
You can select up to five datasets for each SQL workflow. This limit helps to manage the complexity and maintain smooth performance of the tasks.
Enable SQL Explore, Query and Visualize Tool
To run SQL workflows, you need to enable the SQL Explore, Query, and Visualize tool. It's also helpful to turn off the Web Search & Browse tool if you don't expect Athena to use it. This step prevents Athena from using web search instead of analyzing your connected datasets to answer your questions.
Executing Workflows
Sending the Initial Request
To start an SQL workflow, compose your request to Athena and send the message. It's very helpful to provide specific context in your initial message to guide Athena efficiently.
Adding details at the beginning helps Athena understand what you're looking for right away. This could include specifying which data sets to use first, what kind of information you're looking for, and any particular nuances related to your request. Providing clarity from the start enhances Athena's ability to deliver accurate and relevant results.
Query Processing
After Athena finishes processing your request, she presents the results either as a table for complex data or a simple text answer for straightforward queries.
If you have follow-up questions or if Athena's initial answer isn't quite right, you can ask for more details or clarity. Athena can handle these additional queries to ensure you get the information you need. This ongoing conversation capability allows for a dynamic and interactive data exploration experience.
Error Handling
When Athena faces an SQL error during a query, she first attempts to fix the issue on her own. This could involve adjusting her query or reviewing the dataset's metadata to find a better approach. If she's not sure what to do, Athena will turn to you for guidance.
When you send a request to Athena via chat for data processing, Athena can query data multiple times, up to about 15 queries at a time, to get the correct result.
If Athena shares debug information and questions with the you. This helps the you understand what went wrong and how they can help Athena proceed. It's part of Athena's self-correction process when handling SQL workflow errors.
Use Tasks
Navigate to Home to see Athena Tasks. Choose between available analysis tasks to start a predefined SQL workflow in Chat.
Frequently Asked Questions
You can select up to five data sets when setting up a chat session for SQL workflows.
To execute SQL workflows, turn on the SQL Explore, Query, and Visualize tool and turn off the Web Browse tool to avoid confusion.
When Athena encounters an error, she tries to correct her query or reviews the data set's metadata. If needed, she will provide debug information and questions to the user.
Chat sessions, including those for SQL workflows, are saved permanently. You can easily access previous chats from the left panel.