pyarrow_parquet_file_read_write_quickstart.py

1import pyarrow.parquet as pq
2import pyarrow as pa
3import pandas as pd
4
5# Create a sample dataset using a Pandas DataFrame
6df = pd.DataFrame({'one': [-1, 1, 2.5],
7                   'two': ['foo', 'bar', 'baz'],
8                   'three': [True, False, True]},
9                   index=list('abc'))
10
11# Convert the DataFrame to a PyArrow Table
12table = pa.Table.from_pandas(df)
13
14# Write the Table to a Parquet file
15pq.write_table(table, 'example.parquet')
16
17# Read the Parquet file back into a PyArrow Table
18table2 = pq.read_table('example.parquet')
19
20# Convert the Table back to a Pandas DataFrame
21df2 = table2.to_pandas()
22
23print(df2)