I am a python.new who need some help in the following question:

I got a dataframe like this.

df:

...2000 rows

df = pd.DataFrame(data=[[181.5,'cm'],
               [72.5,'inches'],
               [168.0,'cm'],
               ['NaN','NaN']],
         columns = ['height','unit'],
         index=[1,2,3,4])

I want to unify the unit to "cm", and make corresponding changes to height, and keep the 'NaN's.

CodePudding user response：

Use a dictionary to map conversion factors and use indexing to update the values/units:

# ensure real NaNs:
df = df.replace('NaN', np.nan)

# set up dictionary of conversion factors
d = {'cm': 1, 'inches': 2.54}

# map converted heights
df['height'] = df['height'].mul(df['unit'].map(d))
# update units
df.loc[df['unit'].isin(d), 'unit'] = 'cm'

output:

   height unit
1  181.50   cm
2  184.15   cm
3  168.00   cm
4     NaN  NaN

handling unknown units

if you want to handle the case of values for which units are unknown and leave them unchanged, use map(lambda x: d.get(x, 1)) instead of map

CodePudding user response：

Adjusted from this solution using a mask:

mask = (df['unit'] == 'inches')
df_inches = df[mask]
df.loc[mask, 'height'] = df_inches['height'] * 2.54
df.loc[mask, 'unit'] = 'cm'

print(df)

Output:

   height unit
1   181.5   cm
2  184.15   cm
3   168.0   cm
4     NaN  NaN