I’ve faced this issue too. GridDB CE doesn’t support GROUP BY or multi-entity time-window aggregations (like TIME_SLICE) within a single TimeSeries container that holds multiple device IDs. TIME_SLICED aggregation works only on single-entity TimeSeries containers without logical partitioning like deviceId.

The recommended approach in GridDB is to create a separate TimeSeries container per device. This allows you to leverage INTERPOLATE and TIME_SLICED queries efficiently. If managing thousands of containers is a concern, consider automating container management, but unfortunately, CE doesn’t support per-device aggregation in a shared container directly via query