df['full_name'] = df['first'] + '' + df['last']

adding columsn

df.drop(columns = ['first', ' last'], inplace = True)

removing column

df[['first', 'last']] = df['full_name'].str.split(' ', expand = True)

mergin two columns as one

df.dropna(axis ='index', how ='all', subset=['last', 'email'])


droping null values with specfiying indexes

df['age'] = df['age'].astype(float)


working with missing values

df['YearsCode'].unique()

getting only unique values

df['date'] = pd.to_datetime(df.date, dayfirst=True)
# df['month'] = df['date'].dt.month
# df['month_year'] = df['full_datetime'].dt.strftime('%m-%Y')
# df['date'] = df['date'].dt.to_period('M')



when working with dates

Numpy

x = np.array([-1,0,1.4],dtype='bool')
y = np.array([-1,0,1.4],dtype='int16')
z = np.array([-1,0,1.4],dtype='float64')

it says boolean works faster

x = np.array([0,0,0],dtype='uint8')
x[0] = 255
x[1] = x[0] + 1
x[2] = x[1] + 1
print(x)

it means uint8 contains only 255 character

x = np.array([-1,0,1,10],dtype='float64')
print( x / 0 )
y = 10 / 0 # core Python


-inf nan inf inf it says when dividing 0 to 0

a = (np.inf == np.inf)
b = (np.nan == np.nan) # Can not compare nan to nan!
c = np.isnan(np.nan)

you can compare inf to inf
but you can not compare nan to nan

a = np.array([1,2,3,4,5],dtype='float64')
# a = np.arange(5,dtype='uint8') + 1
print([type(a_element) for a_element in a])


for using arrange used list comprehension with type and for

obj_explore(a,'public')

exploration

# Assigning elements of an array
z[1,0] = -1
z[2] = [4,5,7] # assign whole row from a list
z[:,0] = np.array([4.2,5.2,6.2,7.2,8.2]) # assign column from nparray
z[:2,1]=9.3
z[3][1]=-2 # note double bracket indexing
print(z)