For each column group of values order by list without dropping duplicates-CodePudding

if i have a dataframe like this one:

level	org
1	A
1	A
2	A
2	A
3	A
3	A
1	B
1	B
2	B
2	B
3	B
3	B
1	C
1	C
2	C
2	C
3	C
3	C

I want to use a list based on unique values of the level column:

levels=[1,2,3]

to order each individual org group and get an output like this preserving the duplicates:

level	org
1	A
2	A
3	A
1	A
2	A
3	A
1	B
2	B
3	B
1	B
2	B
3	B
1	C
2	C
3	C
1	C
2	C
3	C

CodePudding user response：

Create new column with cumcount

df['new'] = df.groupby(['level','org']).cumcount()
out = df.sort_values(['org','new','level'])
Out[71]: 
    level org  new
0       1   A    0
2       2   A    0
4       3   A    0
1       1   A    1
3       2   A    1
5       3   A    1
6       1   B    0
8       2   B    0
10      3   B    0
7       1   B    1
9       2   B    1
11      3   B    1
12      1   C    0
14      2   C    0
16      3   C    0
13      1   C    1
15      2   C    1
17      3   C    1