dataframe from list of `(row, column, value)`?

Question

I have a list of tuples in (row, column, value) format (coming from sql query). Eg

data = [
    (date_1, brand_x, 5),
    (date_1, brand_y, 8),
    (date_2, brand_x, 12)
]

and I want to convert them to a dataframe

index  | brand_x | brand_y
--------------------------
date_1 |   5     |    8
date_2 |  12     |   NaN

I've been trying with Dataframe.from_records/from_dict and different orient values, but to no avail. What is the correct way to do it?

PS: If there's a way to do it by converting the tuples to dicts first, it's just as well

score 3 · Accepted Answer · answered Oct 30 '18 at 15:24

3

import pandas as pd
data = [
    ('date_1', 'brand_x', 5),
    ('date_1', 'brand_y', 8),
    ('date_2', 'brand_x', 12)
]    

pd.DataFrame(data).pivot_table(index=0,columns=1)

answered Oct 30 '18 at 15:24

Rudolf Morkovskyi

869
5
19

3

Also, `pd.DataFrame(data)pivot(index=0,columns=1,values=2).reset_index()` – Space Impact Oct 30 '18 at 15:28
`pivot` is better since you are not aggregating data via a function. – jpp Oct 30 '18 at 15:30

dataframe from list of `(row, column, value)`?

1 Answers1