Synchronization of variables
There are two synchronization options, and different products may require different strategies.
- Remove a variable when it is has no valid data points.
- Remove the granule when an extracted variable has no data points in the granule.
The first is desired for variables that are to be filled in a later version of the processor, the second is required for a product where no valid data may be present in some granules (i.e. CH4 and ALH).
Some more testing is required to avoid index error or shape mismatches when combining the data.
This is controlled by the
#1 Updated by Maarten Sneep about 2 years ago
- Status changed from In Progress to Resolved
- % Done changed from 80 to 100
Adding a check on the number of valid retrievals before reading the data solves a lot of the synchronisation issues, and removes the need to handle NRT and OFFL differently.
Resolved in 0.10.4