Your BUCKETIZE using bucketizer.mean is using the actual values which fall within the buckets. If you wish to have different values, you should add a BUCKETIZE step using bucketizer.last, at the resolution you want (1'?) and then use FILLNEXT/FILLPREVIOUS to propagate the known values to the empty buckets. Then you can compute your mean as above and the values you expect will be taken into account.
Have in mind that bucketizer.mean is only considering the actual values in the bucket. If what you are expecting is rather to do a weighted average of the values over the bucket period at a given resolution then you need to proceed as suggested above at the chosen resolution. On Tuesday, February 18, 2020 at 12:43:30 PM UTC+1, Steven Harding wrote: > > I'm using Warp10 to store data - it's at per-second resolution, but to > keep data sensible, it's being stored by exception - meaning that > recordings are only made when values change. > > The trouble is that I'm then having issues with retrieving data because of > the way various WarpScript functions work. > > I know I can use the mapped version of FETCH to include pre and > post-boundary items (thank goodness!), which allows me to get the data I > need, but I'm discovering issues when it comes time to BUCKETIZE. > > When I BUCKETIZE my sparse data, I get results that don't represent the > real world conditions, unless I'm extraordinarily lucky with my timing. In > the following situation... > > 1:00 20 > 1:20 20 > 1:21 30 > 2:30 30 > 2:31 20 > 2:40 20 > > ...into 5-minute mean increments, more often than not I'll end up with > this... > > 1:00 20 > 1:25 25 > 2:35 25 > 2:40 20 > > Because there's no additional data points in the hour-long gap, the > bucketizing is very misleading. > > Is there any function or combination of functions - something like a > FILLNEXT - that I can use before I bucketize my GTS? Is there a variant of > INTERPOLATE I can use that will interpolate based on the first and last > VALUES of the GTS in my buckets, instead of the calculated bucket total (ie > instead of interpolating the '25' to '25' in the example above, it would > take the '30' from the 1:21 sample in the 1:25 bucket, and interpolate it > out to the '30' in the first sample in the 2:35 bucket? > > This last solution would be the absolute ideal - it would perform a much > more accurate variant of BUCKETIZE on sparse data. > -- You received this message because you are subscribed to the Google Groups "Warp 10 users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/warp10-users/761aaff9-d2a9-4588-a387-db1b06551883%40googlegroups.com.
