How to transform number that include alphabet to a new numerical order in pandas columns?-CodePudding

I am trying to convert some columns that include number and alphabet at the same time. I want to convert to simple numerical value such as int. I don't want to just transform randomly because they are isbn involved. I want to able to match again. I want to transform that matching as below.

d = {'id': [156972, 102154, 214717, 84897, 275220, 165759, 42099, 265749, 130474, 15822], 
      'isbn': ['2842630521', '037570745X','689710879','783547528','786014091','1561581747','553571095','451210220','034540288X','345832710X',]}
df = pd.DataFrame(data=d)

I want to transform this way.

for id
2842630521 as 0
783547528 as 1
.
.
.
130474 as 9
15822 as 10

for isbn
2842630521 as 0
037570745X as 1 # alphabet involved
.
.
.
034540288X as 9 
345832710X as 10

Is there anyway to do this to column?

Optional: I would also like a way to track back if the number 0 in isbn column to 2842630521.

CodePudding user response：

Use:

#random shuffle order of rows
df1 = df.sample(frac=1)

#create mapping dictionary by id
a, b  = pd.factorize(df1['id'])
d = dict(zip(b, a))

#map original column
df['id'] = df['id'].map(d)