ARROW-2677: [Python] Expose Parquet ZSTD compression

Author: Korn, Uwe <Uwe.Korn@blue-yonder.com>

Closes #2120 from xhochy/ARROW-2677 and squashes the following commits:

585a4bf <Korn, Uwe> flake8
c540796 <Korn, Uwe> Mention possible compression options in docstring
68f83bb <Korn, Uwe> ARROW-2677:  Expose Parquet ZSTD compression

Author: xhochy

python/pyarrow/_parquet.pxd

@@ -102,6 +102,7 @@ cdef extern from "parquet/api/schema.h" namespace "parquet" nogil:
         ParquetCompression_LZO" parquet::Compression::LZO"
         ParquetCompression_BROTLI" parquet::Compression::BROTLI"
         ParquetCompression_LZ4" parquet::Compression::LZ4"
+        ParquetCompression_ZSTD" parquet::Compression::ZSTD"
 
     enum ParquetVersion" parquet::ParquetVersion::type":
         ParquetVersion_V1" parquet::ParquetVersion::PARQUET_1_0"

python/pyarrow/_parquet.pyx

@@ -809,7 +809,8 @@ cdef class ParquetReader:
         return array
 
 cdef int check_compression_name(name) except -1:
-    if name.upper() not in ['NONE', 'SNAPPY', 'GZIP', 'LZO', 'BROTLI', 'LZ4']:
+    if name.upper() not in ['NONE', 'SNAPPY', 'GZIP', 'LZO', 'BROTLI', 'LZ4',
+                            'ZSTD']:
         raise ArrowException("Unsupported compression: " + name)
     return 0
 
@@ -826,6 +827,8 @@ cdef ParquetCompression compression_from_name(str name):
         return ParquetCompression_BROTLI
     elif name == "LZ4":
         return ParquetCompression_LZ4
+    elif name == "ZSTD":
+        return ParquetCompression_ZSTD
     else:
         return ParquetCompression_UNCOMPRESSED
 

python/pyarrow/parquet.py

@@ -251,6 +251,7 @@ def _sanitize_table(table, new_schema, flavor):
     Valid values: {None, 'ms', 'us'}
 compression : str or dict
     Specify the compression codec, either on a general basis or per-column.
+    Valid values: {'NONE', 'SNAPPY', 'GZIP', 'LZO', 'BROTLI', 'LZ4', 'ZSTD'}
 flavor : {'spark'}, default None
     Sanitize schema or set other compatibility options for compatibility"""
 

python/pyarrow/tests/test_parquet.py

@@ -501,7 +501,7 @@ def test_pandas_parquet_configuration_options(tmpdir):
         df_read = table_read.to_pandas()
         tm.assert_frame_equal(df, df_read)
 
-    for compression in ['NONE', 'SNAPPY', 'GZIP', 'LZ4']:
+    for compression in ['NONE', 'SNAPPY', 'GZIP', 'LZ4', 'ZSTD']:
         _write_table(arrow_table, filename.strpath,
                      version="2.0",
                      compression=compression)