Merge two time series with different frequencies

KNIME Analytics Platform
1

Hey,

I'm trying to build an easy-to-use import-note for an NetCDF file. This Note shoud be very common,  I want use it for different operations on diffent variables. The importing is coded in an R Snipped and worked well. But the NetCDF file includes variables with different frequencies and I need to join them before I can get them out of the R Snipped. My first idea was to build an big time series with the timesteps from both time series (rows) and the data of both time series (columns). Thus, the low frequency data have zero-values at timesteps from the high frequency variable (not definied at this timesteps). The outcoming matrix is huge, the procedure is extremly memory consuming and inputfiles above 100mb lead to Out-of-Memory-Error (in R). 

Is the any possibility to work with different frequencies or matrices in Knime?

Greetings from Germany

Thilo Hofmann

 

 

2

Hi Thilo, 

Once you have your data in KNIME, I would first try to use the date (and/or time) field extractor nodes.  These will pull out some keys that you can then use to join data from one table to another.  

 

If your data can't be easily joined by hour, or month (for example), then you will need to do some work to find a join key that works.  Calculating this key will probably be the hard part, but if you can post an example I can have a look and see if any suggestions come to mind.