Timescaledb: Resultset downsampling

Created on 25 Mar 2018  路  5Comments  路  Source: timescale/timescaledb

As real time timeseries datasets can be large and visualizing the same on any browser based technology creates a huge load on the client browsers, it would be good to have the possibility of Downsampling feature within the query itself; such as Largest-Triangle-Three Bucket Downsampling or Douglas Peucker Reduction. This will help reducing the output of data in the sql response itself.

community-request enhancement

Most helpful comment

Or just min/max in each bucket, which avoids losing peaks and valleys

All 5 comments

when downsampling, can we choose to fill null with 0 or previous values?

Or just min/max in each bucket, which avoids losing peaks and valleys

Usually, I restrict the sample size by either limiting the boundaries of the time window or bucketing, though it does in fact sound like a potent feature :).

Is this feature available yet?

+1 for supporting LTTB as part of time_bucket or time_bucket_gapfill

Was this page helpful?
0 / 5 - 0 ratings