在熊猫数据框上合并多索引列名

Question

I have a csv file that has its column names separated over various lines, like this:

ID,Flow,Flow
,,Type
1,21.79,1
2,1129.32,0

So the column names are ID , Flow , and FLow Type . When I read this csv as a multi-indexed dataframe:

df = pandas.read_csv(path, header = [0,1])

I end up with

               ID                Flow      
Unnamed: 0_level_1  Unnamed: 1_level_1  Type 
1                    21.79                1                             
2                    1129.32              0

Is there a simple way to merge the column names, so that I end up with a simple column index with their proper names? In the end, I would like to have the equivalent of

pandas.DataFrame({'ID': [1,2], 'Flow': [21.79, 1129.32], 'Flow Type': [1,0]})

Answer 1

This method assumes you know the order to columns and their names, and number of rows you want to skip.

If so, you could

pd.read_csv('temp.csv', skiprows=2, names=['ID', 'Flow', 'FLow Type'])

Here, column names are ['ID', 'Flow', 'FLow Type'] and we are skipping skiprows=2 from the 'temp.csv' file.

That shall read the csv in intended way. However, this method woudn't work if column order and skiprows changes.