There are several places in the code where performance can be improved, sometimes perhaps significantly, by using the right kinds of parallel pandas operations, rather than by traditional iteration. Performance improvements are very helpful during development as they make the test/debug cycle much more effective. Use --timing to identify slow points in aggregate.py and switch them to more effective code.
There are several places in the code where performance can be improved, sometimes perhaps significantly, by using the right kinds of parallel pandas operations, rather than by traditional iteration. Performance improvements are very helpful during development as they make the test/debug cycle much more effective. Use --timing to identify slow points in aggregate.py and switch them to more effective code.