fastparquet_pandas_dataframe_write_and_read_quickstart.py

1import pandas as pd
2import numpy as np
3from fastparquet import write, ParquetFile
4
5# Create a sample dataframe
6df = pd.DataFrame({'a': np.random.randn(20),
7                   'b': [True, False] * 10,
8                   'c': ['apple', 'orange'] * 10})
9
10# Write the dataframe to a parquet file
11write('outfile.parquet', df)
12
13# Read the parquet file back into a dataframe
14pf = ParquetFile('outfile.parquet')
15df2 = pf.to_pandas()
16
17# Alternatively, read with filters (optional)
18# df3 = pf.to_pandas(['a', 'b'], filters=[('a', '>', 0)])