How to remove rows containing nan in .fits file with Python?

Question

I tried to remove rows where nan occurs in an astropy table like so

a[~np.isnan(a).any(axis=1)]

but i get following error-code:

TypeError: ufunc 'isnan' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''

Answer 1

I'm guessing that you want to remove rows where any column in that row has a NaN. That would be like so:

In [17]: t = Table([[1.0, np.nan, 3.0], [10, 20, np.nan]])

In [18]: t
Out[18]: 
<Table length=3>
  col0    col1 
float64 float64
------- -------
    1.0    10.0
    nan    20.0
    3.0     nan

In [19]: bad = np.logical_or.reduce([np.isnan(col) for col in t.itercols()])

In [20]: bad
Out[20]: array([False,  True,  True])

In [21]: t[~bad]
Out[21]: 
<Table length=1>
  col0    col1 
float64 float64
------- -------
    1.0    10.0

Answer 2

What do the NaNs look like in the table/array? The error you are reporting suggests a mixture of numbers and characters. If you are reading a table with pandas you could use pandas.dropna() which worked great for me in the past.