How to melt and unpivot the dataframe?

My initial df looks as follows:

import pandas as pd

columns = ["ID", "Position1", "Position2", "Position3", "Duration1", "Duration2", "Duration3", "Costs1", "Costs2", "Costs3"]
values1 = [1, 1, "null", 2, 10, "null", 20, 500, "null", 1500]
values2 = [2, 2, 3, 1, 55, 44, 22, 1000, 2000, 4000]
df = pd.DataFrame([values1, values2], columns=columns)

My goal is to get a dataframe with the following scheme:

| ID | Position | Duration | Costs |
|----|----------|----------|-------|
| 1  | 1        | 10       | 500   |
| 1  | 2        | 20       | 1500  |
| 2  | 1        | 22       | 4000  |
| 2  | 2        | 55       | 2000  |
| 2  | 3        | 44       | 1000  |

I think a good start is the following, but I am not sure how to proceed:

df_melted = df.melt(id_vars="ID")
df_melted = df_melted[df_melted.value != "null"]
df_melted["variable"] = df_melted["variable"].str.replace("(\d+)", "")

getting:

|    | ID | variable | value |
|---:|---:|---------:|------:|
|  0 |  1 | Position |     1 |
|  1 |  2 | Position |     2 |
|  3 |  2 | Position |     3 |
|  4 |  1 | Position |     2 |
|  5 |  2 | Position |     1 |
|  6 |  1 | Duration |    10 |
|  7 |  2 | Duration |    55 |
|  9 |  2 | Duration |    44 |
| 10 |  1 | Duration |    20 |
| 11 |  2 | Duration |    22 |
| 12 |  1 |    Costs |   500 |
| 13 |  2 |    Costs |  1000 |
| 15 |  2 |    Costs |  2000 |
| 16 |  1 |    Costs |  1500 |
| 17 |  2 |    Costs |  4000 |

>Solution :

Code

use wide_to_long

pd.wide_to_long(df, stubnames=['Position', 'Duration', 'Costs'], i='ID', j='val')\
  .droplevel(1)[lambda x: x.ne('null').any(axis=1)]\
  .sort_values(['ID', 'Position']).reset_index()

output:

    ID  Position    Duration    Costs
0   1   1           10          500
1   1   2           20          1500
2   2   1           22          4000
3   2   2           55          1000
4   2   3           44          2000