[Python] Pandas sort_values() function on two columns isn't sorting the second column? [closed]

I have a dataframe that I'm trying to sort by two columns - first by 'ROI', then by 'campaign_id'.

The code is as follows:

final_table.sort_values(by = ['ROI', 'campaign_id'], ascending = [False, False])

The resulting dataframe is successfully sorted by ROI, but has this going on:

campaign_id campaign_name ROI
49 Campaign 49 172
15 Campaign 15 172
73 Campaign 73 172

I've checked the datatypes; ROI is a float and campaign_id is an int. When I try to sort by just campaign_id, it sorts by those values just fine. What am I missing here?

EDIT:

Running the code final_table.dtypes gives the following output:

campaign_id int64

campaign_name object

ROI float64

dtype: object

The campaign_id column contains no repeat values - it's an index that increments for each row of the original dataset. Here's some more of the output:

campaign_id campaign_name ROI
31 Campaign 31 176
67 Campaign 67 176
70 Campaign 70 175
49 Campaign 49 172
15 Campaign 15 172
73 Campaign 73 172
68 Campaign 68 170

Continue reading...