Hi all,

I'm writing to introduce a new PR that significantly improves how we handle 
large query results in the IoTDB Python client.

Fetching massive datasets all at once can often lead to memory issues and 
inefficiencies. To solve this, we are introducing a streaming DataFrame 
interface that allows users to fetch results in manageable chunks.

This PR adds `has_next_df()` and `next_df()` methods to `SessionDataSet`, 
making it much easier to iterate through data. We've also implemented internal 
buffering within `IoTDBRpcDataSet` to manage these data chunks smoothly based 
on the specified `fetch_size`. Additionally, the core logic for processing 
results has been refactored to support both this new streaming approach and the 
traditional batch method.

We have updated the example scripts to demonstrate how to use this new API, 
including specific examples for Table Mode usage.

The PR link is: https://github.com/apache/iotdb/pull/17035

Best regards,
--------------------
Yuan Tian

Reply via email to