Fill in a pandas DataFrame bucketing the values of a column of a different frame and preserving the index

Question

I have a DataFrame of the following type:

df = pd.DataFrame({'price':[1,2,3,2,2,3,1,2,1], 'quantity':[10,20,30,10,20,30,20,20,10]})
df
Out[10]: 
   price  quantity
0      1        10
0      2        20
1      3        30
1      2        10
1      4        20
2      3        30
3      1        20
4      2        20
4      1        10

and I want to create a second DateFrame that looks like:

df_bucket = pd.DataFrame(columns=np.arange(0, 5, 1), index=df.index)
     0    1    2    3    4
0  NaN  10   20   NaN  NaN
1  NaN  NaN  10   30   20
2  NaN  NaN  NaN  30   NaN
3  20   NaN  NaN  NaN  NaN
4  10   20   NaN  NaN  NaN

I tried the following, but it is extremely slow and yields just a bunch of nans

df_bucket.loc[df.index][df['price']] = df['quantity']
df_bucket
Out[12]: 
     0    1    2    3    4
0  NaN  NaN  NaN  NaN  NaN
1  NaN  NaN  NaN  NaN  NaN
2  NaN  NaN  NaN  NaN  NaN
3  NaN  NaN  NaN  NaN  NaN
4  NaN  NaN  NaN  NaN  NaN

I know I can do this in a for loop, but I am sure it's going to take ages, do you know of any faster way to accomplish this?

To give some context, these are order book data indexed by mostly unique timestamps. I would like to reorganise the DataFrame with a column per price level and preserve the index, effectively a very inefficient, but convenient, way of organising the data. The DataFrame has few hundred thousands rows and this is why I need a more efficient way than looping over the rows.

Answer 1

df = pd.DataFrame({'ind':[0,0,1,1,1,2,3,4,4],
                    'price':[1,2,3,2,2,3,1,2,1], 
                   'quantity':[10,20,30,10,20,30,20,20,10]})
df.pivot_table(index=df.ind,columns='price',values='quantity')
df

Output :

 price   1       2      3
    ind         
    0   10.0    20.0    NaN
    1   NaN     15.0    30.0
    2   NaN     NaN     30.0
    3   20.0    NaN     NaN
    4   10.0    20.0    NaN