Dealing with numeric range-like data in a dataframe

Question

I have a dataframe that has values like 1,362 - 2,037, first, how can I clean them? like getting the mean of these two numbers (10 - 20 ➡ 15), normalizing them or anything else? second, how can I remove the comma in 1,362 to 1362 in all rows?

Currently Have:

col1	col2	col3
Yes	579	402 - 1,120
No	1,082	1,361 - 2,037

Expected Output:

(mean, normalization, etc.)

col1	col2	col3
Yes	579	761
No	1082	1699

shape of dataframe:

df = pd.read_csv("/content/cleanedNASA.csv")
df.describe
...
> [31 rows x 9 columns]

thanks.

Answer 1

df['y'] = pd.to_numeric(df['x']
    .str.replace('[^0-9\- ]', '') # 1. remove non-(digit/dash/space) characters
    .str.split(' - ')             # 2. split by ' - '
    .explode()                    # 3. explode the lists
).mean(level=0)                   # 4. calculate average by index

df

Output:

               x     y
0    402 - 1,120   761
1            674   674
2          3,196  3196
3          1,304  1304
4            853   853
5  1,361 - 2,037  1699