Open tmcphillips opened 6 years ago
A new data cube is being created. ETA is currently unknown but probably done by Friday.
The corrected data cubes were finished on Friday (7/13) but had poor performance when graphing due to the band interleave vs pixel interleave. Converting the data cube to pixel interleave improved graphing time from 40-60 seconds to tenths of a second.
But a further wrinkle came up. the TMAX variable was created with a different tiling option. Converting the TMAX variable to pixel interleave managed 10% in 48-hours vs 2 hours for the other variables.
Re-processing the TMAX variable now, ETA Wednesday (7/18) evening.
It looks like the 120 GB data cubes used by the timeseries service for the prism data set variables are entirely zeros. Running gdallocationinfo at various positions returns zeros for all bands, and running gdalinfo -stats on each file confirms that the min and max values for each band is zero. This is true both for the precipitation variable and for the minimum temperature variable.
Also, the data cube for maximum temperature is just a symbolic link to the minimum temperature data cube.