Closed sdtaylor closed 6 years ago
doesn't help, an sqlite file is actually bigger than the flat csv. but it could still help speed things up
Instead of sqlite I summarized all the bootstrapped models in the predictions code. this brings predictions.csv
down to ~10mb
https://github.com/sdtaylor/phenology_dataset_study/commit/fabbb430be4c1516c78c7155ab73c27e107725ae
the results file containing all the predictions is 1.5gb now and just too big. the model parameter file at 35mb could probably be put into one too