SNR flag about clustr HOT 5 CLOSED

There may be NaN's or empty values in the SNR column. I'd print out the column to inspect it.

from clustr.

It does, I have found a few. Jose suggested removing the NaNs before making the flag cuts, but I am having some issues getting that to work. This is the error I'm getting for that:
IndexError: boolean index did not match indexed array along dimension 0; dimension is 205 but corresponding boolean dimension is 1092
I'm trying to figure out if I should try to change the NaNs in the catalog, or keep trying to find a way to remove them before the SNR cut.

from clustr.

Ok I just checked - the NaN check was only in place for get_data(): https://github.com/sweverett/CluStR/blob/master/clustr.py#L267

The flag cuts don't currently have this check. So if you are creating a "mask" which contains 1 for things that you want to cut, then it would look something like this:

mask |= (flag_condition) | (np.isnan(catalog[flag_column]))

What this does is take the existing mask and add new mask values for this specific flag (mask |= corresponds to mask = mask | ...) . Instead of just having the flag condition that you usually have, you also look to see if there are NaN's for flag values. This says a cluster should be flagged if the condition is met or if there is a NaN.

from clustr.

I don't make the mask until after the greater than statement, so I get the same error when I put that in. I believe it is working and just ignoring the NaNs. Should I just suppress the error and move on for now? or try to figure it out?

from clustr.

Works for version 2c3f335 clustr.py.

The issue came from using NaNs with comparison operators.

NaNs are dealt with by equating them to dummy variables then comparing them to cut_off_value.

from clustr.