Hello,
I use Potsgres 13 with timescale 2.15.3. I have a table partitioned to 170 chunks.
I want to get records from time range that is contained by only one chunk.
When i get them directly from chunk (select * from _timescaledb_internal._hyper_19_4440_chunk where…) execution time is approximately 100-200 ms less than when getting from table (select * from table where…). Number of returned rows is the same, execution plan, used index. Timescale is tuned. What might be a reason? Is there a way to reindex some internal tables of timescale to decrease execution of queries from hypertable?
Can you post your query more in detail,
Share the specific WHERE condition used in both queries.
Provide information on whether any compression policies or segmenting strategies are set for the hypertable.
The best possible way to find is by Exaplin query:
You can use the EXPLAIN command to generate the execution plans. Look for any differences that might indicate inefficiencies in how the hypertable query is processed.