Support pl.Array when writing parquet files

[mkleinbort-ic]

Description

The Array datatype is not yet supported in writing to parquet

import polars as pl 

df = pl.DataFrame({
    'x':[[1,2,3]]
})

df.cast({'x':pl.Array(pl.Int64, width=3)}).write_parquet('test.parquet')

>>> PanicException: not yet implemented: Writing FixedSizeList to parquet not yet implemented

[deanm0000]

I added "bug" since it shouldn't raise PanicException.

Interesting that if you use pyarrow to write it that polars can then read back the array containing parquet.

[franz101]

import polars as pl
import numpy as np
import pyarrow.parquet as pq

df = pl.DataFrame(
    {
        "arr": np.random.rand(100_000, 512),
        "arr2": np.random.rand(100_000, 512),
    },
    schema={
        "arr": pl.Array(inner=pl.Float32, width=512),
        "arr2": pl.Array(inner=pl.Float32, width=512),
    },
)
pt = df.to_arrow()
pq.write_table(pt, 'example_table.parquet')

Simple way using parquet as a workaround