nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Oracle Communities

Oracle Analytics and AI

Child Item

Agent Flow

User_3XIO0

I have a simple AIDP workspace setup in Chicago. Within it I have a delta table stored at test_catalog.test_schema.clean_sales_data. This table contains some columns which are all strings and a few rows of data.

I can query it using a notebook without any issues.

I have set up an agent flow. That contains Chat Trigger → Agent → SQL Tool

The SQL tool is pointing at the catalog and schema and contains a simple description and query

Tool Name: List sales seasons
Description: Use this tool to list the seasons that sales are catagorised by.

SELECT distinct season
FROM clean_store_sales

When I test it I get an error Failed to fetch results, internal server error

When I look at the compute all the clusters are up, the master, spark cluster and ai cluster.

I don't see where the logs are stored. Has anybody else encountered this issue or tried working with the agent flows?

Find more posts tagged with

AIDP Workbench

agentflow

Accepted answers

User_I64N9

Hi @User_3XIO0

Currently SQLTool in Agent Flow doesn't support querying standard catalog (This feature is in roadmap). We can query external catalog over oracle database though. For now, you'll have to write the content of the delta table to a table in external catalog. If you have an upstream pipeline writing data as a delta table and if this delta table has to be queried only via the SQLTool, you may consider switching the destination to a table in external catalog to avoid data duplication.

If the volume of delta table is low, the other option could be to keep the content as a file in workspace or volume and then read it via python using agent in high-code.

All comments

User_I64N9

Hi @User_3XIO0

If the volume of delta table is low, the other option could be to keep the content as a file in workspace or volume and then read it via python using agent in high-code.

User_3XIO0

Thank you @User_I64N9 - that is good to know.

Do you know where that info is documented please? I couldn't find anything in the docs explaining.

User_I64N9

Hi @User_3XIO0 Since these features are currently in LA (Limited Availability), the documentation isn't available in public. It will be released post GA.

User_3XIO0

OK, I understand - thanks very much for the info.

RVohra

Hi @User_3XIO0 , Quick observation, your delta table is stored at test_catalog.test_schema.clean_sales_data.

But your query is from clean_store_sales but table name mentioned above is clean_sales_data.

SELECT distinct season FROM clean_store_sales

Hope you have looked at it. Thanks

Bhaskar Konar

Very helpful information.

Thanks for sharing @User_I64N9!