Varying points in Windows

https://learninghub.kx.com/forums/topic/varying-points-in-windows

In the recent time series similarity search webinar, it is mentioned that windows must be identical time frame with varying number of points in each window, however in the example schema, the dimensions of the windows is set to 1000 data points with no regard to the time frame. Can anyone provide clarity on this?

Hi Trinity,


For Transformed TSS, the 'dims' parameter is not required within the vectorIndex portion of the schema. You should be able to remove that with no problems. As you mentioned, when creating the sliding time series windows to be ingested into KDB.AI, ensure each window is of the same time frame.

Thank you,

Ryan