How to read parquet files into array of python dictionary?

Question

Is there a way to convert parquet files into an array of python dictionaries where the keys are the columns?

import pyarrow.parquet as pq

Does that library natively support that feature?

Answer 1

Use to_pylist

import pandas as pd
import pyarrow.parquet as pq

df = pd.concat(
    [
        pd.Series([1, 2, 3], name='col1'), 
        pd.Series(["abc", "def", "ghi"], name='col2')
    ],
    axis=1
)
df.to_parquet('df.parquet')
pq.read_table('df.parquet').to_pydict()
>>> {'col1': [1, 2, 3], 'col2': ['abc', 'def', 'ghi']}