Python pandas dataframe convert values without decimals

In my dataframe, I have changed the NaN to 0 with the following code

df5.fillna(0, inplace=True)

However, I get the value "0.0" instead of "0". I have >150 columns in this dataframe, some need the decimals, and the ones with converted NaN values must be without decimals. How can I get that. My dataframe is for example as follows:

           genome  contig  genes    SCM  SCM/genes  TrfA_1__CP11611  \
source                                                                
20900_48    20900      48      1    0.0       0.00              NaN   
20900_37    20900      37    130  103.0       0.79              Nan

I get:

           genome  contig  genes    SCM  SCM/genes  TrfA_1__CP11611  \
source                                                                
20900_48    20900      48      1    0.0       0.00              0.0   
20900_37    20900      37    130  103.0       0.79              0.0

I need only "NaN" changed to "0" without affecting for example column SCM/genes. It is no option to use a code with columnnames, since I have >150 columns with NaN in this dataframe.

Thanks

Solution

I think you first filter the cols that contain NaN, then convert these:

In [26]:
nan_cols = df.columns[df.isnull().any(axis=0)]
nan_cols

Out[26]:
Index(['TrfA_1__CP11611'], dtype='object')

In [27]:
for col in nan_cols:
    df[col] = df[col].fillna(0).astype(int)
df

Out[27]:
          enome  contig  genes    SCM  SCM/genes  TrfA_1__CP11611
source                                                           
20900_48  20900      48      1    0.0       0.00                0
20900_37  20900      37    130  103.0       0.79                0

So this first looks for NaN present in any rows and makes a list of the cols, you can then iterate over the cols and call fillna and cast the dtype using astype so that you preserve/convert the dtype.