Shift plugin creates NaNs, low pass filter does not handle NaNs (or so it seems).

[wowczarek]

Hi kst team,

I am dealing with the following issue:

I am doing some phase / frequency stability analysis and I need first and second differences between consecutive data points. Being too lazy to write a plugin or script it, I do a left shift on the vector, subtract, shift again and subtract. The problem is that the shift does not shorten the vector, it slides in NaNs. I needed to run lpf on the result, and all I get is [n x NaN]. My only guess is that those NaNs in the end of the vector cause lpf to produce this result. Should it not be able to ignore NaNs? Likewise, is there a good reason why the shift plugin inserts NaNs? Pehaps an option to do zero-fill or shorten the buffer?

The source says "pad beginning with zeros", but then fills it with NaNs.

...I wanted to add that I've been using kst on a daily basis for many years for realtime analysis of data coming in from various measurement systems. Having the ability to work on live files is a huge help, and I know I can throw any amount of data at it and it still flies. People who have not heard of kst truly don't know what they are missing!

Cheers,
Wojciech