Reading and Writing the Apache ORC Format

The Apache ORC project provides a standardized open-source columnar storage format for use in data analysis systems. It was created originally for use in Apache Hadoop with systems like Apache Drill, Apache Hive, Apache Impala, and Apache Spark adopting it as a shared standard for high performance data IO.

Apache Arrow is an ideal in-memory representation layer for data that is being read or written with ORC files.

Obtaining pyarrow with ORC Support

If you installed pyarrow with pip or conda, it should be built with ORC support bundled:

>>> from pyarrow import orc

If you are building pyarrow from source, you must use -DARROW_ORC=ON when compiling the C++ libraries and enable the ORC extensions when building pyarrow. See the Python Development page for more details.

Reading and Writing Single Files

The functions read_table() and write_table() read and write the pyarrow.Table object, respectively.

Let’s look at a simple table:

>>> import numpy as np
>>> import pyarrow as pa

>>> table = pa.table(
...     {
...         'one': [-1, np.nan, 2.5],
...         'two': ['foo', 'bar', 'baz'],
...         'three': [True, False, True]
...     }
... )

We write this to ORC format with write_table:

>>> from pyarrow import orc
>>> orc.write_table(table, 'example.orc')

This creates a single ORC file. In practice, an ORC dataset may consist of many files in many directories. We can read a single file back with read_table:

>>> table2 = orc.read_table('example.orc')

You can pass a subset of columns to read, which can be much faster than reading the whole file (due to the columnar layout):

>>> orc.read_table('example.orc', columns=['one', 'three'])
pyarrow.Table
one: double
three: bool
----
one: [[-1,nan,2.5]]
three: [[true,false,true]]

We need not use a string to specify the origin of the file. It can be any of:

  • A file path as a string

  • A Python file object

  • A pathlib.Path object

  • A NativeFile from PyArrow

In general, a Python file object will have the worst read performance, while a string file path or an instance of NativeFile (especially memory maps) will perform the best.

We can also read partitioned datasets with multiple ORC files through the pyarrow.dataset interface.

ORC file writing options

write_table() has a number of options to control various settings when writing an ORC file.

  • file_version, the ORC format version to use. '0.11' ensures compatibility with older readers, while '0.12' is the newer one.

  • stripe_size, to control the approximate size of data within a column stripe. This currently defaults to 64MB.

See the write_table() docstring for more details.

Finer-grained Reading and Writing

read_table uses the ORCFile class, which has other features:

>>> orc_file = orc.ORCFile('example.orc')
>>> orc_file.metadata

-- metadata --
>>> orc_file.schema
one: double
two: string
three: bool
>>> orc_file.nrows
3

See the ORCFile docstring for more details.

As you can learn more in the Apache ORC format, an ORC file consists of multiple stripes. read_table will read all of the stripes and concatenate them into a single table. You can read individual stripes with read_stripe:

>>> orc_file.nstripes
1
>>> orc_file.read_stripe(0)
pyarrow.RecordBatch
one: double
two: string
three: bool

We can write an ORC file using ORCWriter:

>>> with orc.ORCWriter('example2.orc') as writer:
...     writer.write(table)

Compression

The data pages within a column in a row group can be compressed after the encoding passes (dictionary, RLE encoding). In PyArrow we don’t use compression by default, but Snappy, ZSTD, Gzip/Zlib, and LZ4 are also supported:

>>> orc.write_table(table, where, compression='uncompressed')
>>> orc.write_table(table, where, compression='gzip')
>>> orc.write_table(table, where, compression='zstd')
>>> orc.write_table(table, where, compression='snappy')

Snappy generally results in better performance, while Gzip may yield smaller files.

Reading from cloud storage

In addition to local files, pyarrow supports other filesystems, such as cloud filesystems, through the filesystem keyword:

>>> from pyarrow import fs

>>> s3  = fs.S3FileSystem(region="us-east-2")
>>> table = orc.read_table("bucket/object/key/prefix", filesystem=s3)