add empty test0.parquet + inspect test

tests/test0_inspect.txt

@@ -0,0 +1,56 @@
+
+############ file meta data ############
+created_by: parquet-cpp-arrow version 14.0.2
+num_columns: 4
+num_rows: 0
+num_row_groups: 1
+format_version: 2.6
+serialized_size: 2398
+
+
+############ Columns ############
+a
+b
+c
+d
+
+############ Column(a) ############
+name: a
+path: a
+max_definition_level: 1
+max_repetition_level: 0
+physical_type: DOUBLE
+logical_type: None
+converted_type (legacy): NONE
+compression: SNAPPY (space_saved: -7%)
+
+############ Column(b) ############
+name: b
+path: b
+max_definition_level: 1
+max_repetition_level: 0
+physical_type: INT32
+logical_type: Null
+converted_type (legacy): NONE
+compression: SNAPPY (space_saved: -7%)
+
+############ Column(c) ############
+name: c
+path: c
+max_definition_level: 1
+max_repetition_level: 0
+physical_type: INT64
+logical_type: None
+converted_type (legacy): NONE
+compression: SNAPPY (space_saved: -7%)
+
+############ Column(d) ############
+name: d
+path: d
+max_definition_level: 1
+max_repetition_level: 0
+physical_type: BOOLEAN
+logical_type: None
+converted_type (legacy): NONE
+compression: SNAPPY (space_saved: N/A)
+

tests/test_parquet.py

@@ -1,3 +1,8 @@
+from os import path
+from os.path import dirname
+
+from subprocess import check_output
+
 from parquet_tools.parquet.reader import get_filemetadata
 from parquet_tools.gen_py.parquet.ttypes import (FileMetaData, SchemaElement, LogicalType, StringType, RowGroup, ColumnMetaData, ColumnChunk,
                                                  Statistics, PageEncodingStats, KeyValue)
@@ -216,3 +221,12 @@ def test_key_value_metadata(self, fmd):
                 key='ARROW:schema',
                 value='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')
         ]
+
+    def test_inspect(self):
+        tests_dir = dirname(__file__)
+        test0_parquet_path = path.join(tests_dir, 'test0.parquet')
+        actual = check_output(['parquet-tools', 'inspect', test0_parquet_path]).decode()
+        expected_path = path.join(tests_dir, 'test0_inspect.txt')
+        with open(expected_path, 'r') as f:
+            expected = f.read()
+        assert actual == expected

tests/test_parquets.py

@@ -17,6 +17,8 @@ def write_test_dataframes():
     tests_dir = dirname(__file__)
     df.to_parquet(f'{tests_dir}/test1.parquet')
     df.to_parquet(f'{tests_dir}/test2.parquet')
+    df0 = pd.DataFrame({ 'a': [], 'b': [], 'c': [], 'd': [], }).astype({ 'a': float, 'b': str, 'c': int, 'd': bool })
+    df0.to_parquet(f'{tests_dir}/test0.parquet')
 
 
 if __name__ == '__main__':

tests/test_util.py

@@ -19,6 +19,7 @@ def test_is_wildcard(self, pf, expected):
     @pytest.mark.parametrize('pf, expected', [
         (
             LocalParquetFile(path='./tests/*.parquet'), [
+                LocalParquetFile('./tests/test0.parquet'),
                 LocalParquetFile('./tests/test1.parquet'),
                 LocalParquetFile('./tests/test2.parquet')
             ]
@@ -82,11 +83,13 @@ def test_multiple_localfile(self):
             LocalParquetFile(path='./tests/*.parquet'),
         )
 
-        assert len(actual) == 2
+        assert len(actual) == 3
         assert isinstance(actual[0], LocalParquetFile)
         assert isinstance(actual[1], LocalParquetFile)
+        assert isinstance(actual[1], LocalParquetFile)
 
         assert {a.path for a in actual} == {
+            './tests/test0.parquet',
             './tests/test1.parquet',
             './tests/test2.parquet',
         }