I’m working with GridDB to store and query time-series data. The data represents sensor readings taken every minute and is stored in a container with the following structure:

CREATE TABLE sensor_readings (
    timestamp TIMESTAMP,
    sensor_id INT,
    reading DOUBLE
);

I need to write a SQL query that calculates a rolling sum of the reading values over a 30-minute window for each sensor. The challenge is that the dataset contains millions of rows, and I want to optimize the query to ensure it runs efficiently on GridDB.

Specifically, I want the following:

1. Ensure that the query can be run efficiently even as the dataset grows, ideally using any GridDB-specific features like partitioning, window functions, or aggregation optimizations.
2. Calculate the rolling sum of reading for each sensor_id over the last 30 minutes.

Here’s what I’ve tried so far:

SELECT sensor_id, timestamp, 
       SUM(reading) OVER (PARTITION BY sensor_id 
                          ORDER BY timestamp 
                          RANGE BETWEEN INTERVAL '30 MINUTE' PRECEDING AND CURRENT ROW) AS rolling_sum
FROM sensor_readings;

However, this query is not performing well, especially on larger datasets. I suspect it’s because GridDB is not handling the rolling sum efficiently across the large time series.

Can anyone suggest how to optimize this query, possibly by using GridDB-specific indexing, partitioning strategies, or SQL techniques for better performance? Also, are there any GridDB-specific features that could help me with time-based rolling aggregations?