Using a value of 0 will cause problems if we're using a product-based stacking approach since it will zero the entire stack. This should be really obvious in the stack function, but is still non-ideal. One solution would be to use fill_value=np.nan instead in conjunction with np.nansum() and np.nanprod() but the presence of NaNs in a Stream is likely to crash ObsPy functions such as filtering, etc.
Original report by me.
Using a value of 0 will cause problems if we're using a product-based stacking approach since it will zero the entire stack. This should be really obvious in the stack function, but is still non-ideal. One solution would be to use
fill_value=np.nan
instead in conjunction withnp.nansum()
andnp.nanprod()
but the presence of NaNs in a Stream is likely to crash ObsPy functions such as filtering, etc.