deltacat 2.0.0b11__py3-none-any.whl → 2.0.0.post1__py3-none-any.whl

deltacat/tests/utils/test_pyarrow.py

@@ -1,27 +1,50 @@
 from unittest import TestCase
 from deltacat.utils.pyarrow import (
-    s3_partial_parquet_file_to_table,
+    partial_parquet_file_to_table,
     pyarrow_read_csv,
+    ContentTypeValidationError,
     content_type_to_reader_kwargs,
     _add_column_kwargs,
-    s3_file_to_table,
+    file_to_table,
+    file_to_parquet,
+    table_to_file,
     ReadKwargsProviderPyArrowSchemaOverride,
+    ReadKwargsProviderPyArrowCsvPureUtf8,
+    RAISE_ON_DECIMAL_OVERFLOW,
     RAISE_ON_EMPTY_CSV_KWARG,
 )
+import decimal
 from deltacat.types.media import ContentEncoding, ContentType
 from deltacat.types.partial_download import PartialParquetParameters
 from pyarrow.parquet import ParquetFile
+import tempfile
 import pyarrow as pa
+from pyarrow import csv as pacsv
+import fsspec
+import gzip
+import json
+from pyarrow import (
+    feather as paf,
+    parquet as papq,
+    orc as paorc,
+)
 
 PARQUET_FILE_PATH = "deltacat/tests/utils/data/test_file.parquet"
+PARQUET_GZIP_COMPRESSED_FILE_PATH = "deltacat/tests/utils/data/test_file.parquet.gz"
 EMPTY_UTSV_PATH = "deltacat/tests/utils/data/empty.csv"
 NON_EMPTY_VALID_UTSV_PATH = "deltacat/tests/utils/data/non_empty_valid.csv"
+OVERFLOWING_DECIMAL_PRECISION_UTSV_PATH = (
+    "deltacat/tests/utils/data/overflowing_decimal_precision.csv"
+)
+OVERFLOWING_DECIMAL_SCALE_UTSV_PATH = (
+    "deltacat/tests/utils/data/overflowing_decimal_scale.csv"
+)
 GZIP_COMPRESSED_FILE_UTSV_PATH = "deltacat/tests/utils/data/non_empty_compressed.gz"
 BZ2_COMPRESSED_FILE_UTSV_PATH = "deltacat/tests/utils/data/non_empty_compressed.bz2"
 
 
-class TestS3PartialParquetFileToTable(TestCase):
-    def test_s3_partial_parquet_file_to_table_sanity(self):
+class TestPartialParquetFileToTable(TestCase):
+    def test_partial_parquet_file_to_table_sanity(self):
 
         pq_file = ParquetFile(PARQUET_FILE_PATH)
         partial_parquet_params = PartialParquetParameters.of(
@@ -35,7 +58,7 @@ class TestS3PartialParquetFileToTable(TestCase):
         # only first row group to be downloaded
         partial_parquet_params.row_groups_to_download.pop()
 
-        result = s3_partial_parquet_file_to_table(
+        result = partial_parquet_file_to_table(
             PARQUET_FILE_PATH,
             include_columns=["n_legs"],
             content_encoding=ContentEncoding.IDENTITY.value,
@@ -46,7 +69,7 @@ class TestS3PartialParquetFileToTable(TestCase):
         self.assertEqual(len(result), 3)
         self.assertEqual(len(result.columns), 1)
 
-    def test_s3_partial_parquet_file_to_table_when_schema_passed(self):
+    def test_partial_parquet_file_to_table_when_schema_passed(self):
 
         pq_file = ParquetFile(PARQUET_FILE_PATH)
         partial_parquet_params = PartialParquetParameters.of(
@@ -66,7 +89,7 @@ class TestS3PartialParquetFileToTable(TestCase):
 
         pa_kwargs_provider = lambda content_type, kwargs: {"schema": schema}
 
-        result = s3_partial_parquet_file_to_table(
+        result = partial_parquet_file_to_table(
             PARQUET_FILE_PATH,
             ContentType.PARQUET.value,
             ContentEncoding.IDENTITY.value,
@@ -85,7 +108,7 @@ class TestS3PartialParquetFileToTable(TestCase):
         self.assertEqual(result_schema.field(2).type, "int64")
         self.assertEqual(result_schema.field(2).name, "MISSING")
 
-    def test_s3_partial_parquet_file_to_table_when_schema_missing_columns(self):
+    def test_partial_parquet_file_to_table_when_schema_missing_columns(self):
 
         pq_file = ParquetFile(PARQUET_FILE_PATH)
         partial_parquet_params = PartialParquetParameters.of(
@@ -105,7 +128,7 @@ class TestS3PartialParquetFileToTable(TestCase):
 
         pa_kwargs_provider = lambda content_type, kwargs: {"schema": schema}
 
-        result = s3_partial_parquet_file_to_table(
+        result = partial_parquet_file_to_table(
             PARQUET_FILE_PATH,
             ContentType.PARQUET.value,
             ContentEncoding.IDENTITY.value,
@@ -122,7 +145,7 @@ class TestS3PartialParquetFileToTable(TestCase):
         self.assertEqual(result_schema.field(0).type, "int64")
         self.assertEqual(result_schema.field(0).name, "MISSING")
 
-    def test_s3_partial_parquet_file_to_table_when_schema_passed_with_include_columns(
+    def test_partial_parquet_file_to_table_when_schema_passed_with_include_columns(
         self,
     ):
 
@@ -139,11 +162,11 @@ class TestS3PartialParquetFileToTable(TestCase):
 
         pa_kwargs_provider = lambda content_type, kwargs: {"schema": schema}
 
-        result = s3_partial_parquet_file_to_table(
+        result = partial_parquet_file_to_table(
             PARQUET_FILE_PATH,
             ContentType.PARQUET.value,
             ContentEncoding.IDENTITY.value,
-            ["n_legs", "animal"],
+            column_names=["n_legs", "animal"],
             pa_read_func_kwargs_provider=pa_kwargs_provider,
             partial_file_download_params=partial_parquet_params,
         )
@@ -155,7 +178,7 @@ class TestS3PartialParquetFileToTable(TestCase):
         self.assertEqual(result_schema.field(0).type, "string")
         self.assertEqual(result_schema.field(0).name, "n_legs")  # order doesn't change
 
-    def test_s3_partial_parquet_file_to_table_when_multiple_row_groups(self):
+    def test_partial_parquet_file_to_table_when_multiple_row_groups(self):
 
         pq_file = ParquetFile(PARQUET_FILE_PATH)
         partial_parquet_params = PartialParquetParameters.of(
@@ -166,7 +189,7 @@ class TestS3PartialParquetFileToTable(TestCase):
             partial_parquet_params.num_row_groups, 2, "test_file.parquet has changed."
         )
 
-        result = s3_partial_parquet_file_to_table(
+        result = partial_parquet_file_to_table(
             PARQUET_FILE_PATH,
             content_encoding=ContentEncoding.IDENTITY.value,
             content_type=ContentType.PARQUET.value,
@@ -407,130 +430,1388 @@ class TestReadCSV(TestCase):
             ),
         )
 
+    def test_read_csv_when_decimal_precision_overflows_and_raise_kwarg_specified(self):
+        schema = pa.schema(
+            [("is_active", pa.string()), ("decimal_value", pa.decimal128(4, 2))]
+        )
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        _add_column_kwargs(
+            ContentType.UNESCAPED_TSV.value,
+            ["is_active", "decimal_value"],
+            ["is_active", "decimal_value"],
+            kwargs,
+        )
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
 
-class TestS3FileToTable(TestCase):
-    def test_s3_file_to_table_identity_sanity(self):
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
+        self.assertRaises(
+            pa.lib.ArrowInvalid,
+            lambda: pyarrow_read_csv(
+                OVERFLOWING_DECIMAL_PRECISION_UTSV_PATH,
+                **{**kwargs, RAISE_ON_DECIMAL_OVERFLOW: True},
+            ),
+        )
 
+    def test_read_csv_when_decimal_precision_overflows_sanity(self):
         schema = pa.schema(
-            [("is_active", pa.string()), ("ship_datetime_utc", pa.timestamp("us"))]
+            [("is_active", pa.string()), ("decimal_value", pa.decimal128(4, 2))]
+        )
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        _add_column_kwargs(
+            ContentType.UNESCAPED_TSV.value,
+            ["is_active", "decimal_value"],
+            ["is_active", "decimal_value"],
+            kwargs,
         )
 
-        result = s3_file_to_table(
-            NON_EMPTY_VALID_UTSV_PATH,
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
+
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
+
+        self.assertRaises(
+            pa.lib.ArrowInvalid,
+            lambda: pyarrow_read_csv(OVERFLOWING_DECIMAL_PRECISION_UTSV_PATH, **kwargs),
+        )
+
+    def test_read_csv_when_decimal_scale_overflows_and_raise_kwarg_specified(self):
+        schema = pa.schema(
+            [("is_active", pa.string()), ("decimal_value", pa.decimal128(20, 2))]
+        )
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        _add_column_kwargs(
             ContentType.UNESCAPED_TSV.value,
-            ContentEncoding.IDENTITY.value,
-            ["is_active", "ship_datetime_utc"],
-            None,
-            pa_read_func_kwargs_provider=ReadKwargsProviderPyArrowSchemaOverride(
-                schema=schema
+            ["is_active", "decimal_value"],
+            ["is_active", "decimal_value"],
+            kwargs,
+        )
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
+
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
+
+        self.assertRaises(
+            pa.lib.ArrowInvalid,
+            lambda: pyarrow_read_csv(
+                OVERFLOWING_DECIMAL_SCALE_UTSV_PATH,
+                **{**kwargs, RAISE_ON_DECIMAL_OVERFLOW: True},
             ),
         )
 
+    def test_read_csv_when_decimal_scale_overflows_sanity(self):
+        schema = pa.schema(
+            [("is_active", pa.string()), ("decimal_value", pa.decimal128(20, 2))]
+        )
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        _add_column_kwargs(
+            ContentType.UNESCAPED_TSV.value,
+            ["is_active", "decimal_value"],
+            ["is_active", "decimal_value"],
+            kwargs,
+        )
+
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
+
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
+
+        result = pyarrow_read_csv(OVERFLOWING_DECIMAL_SCALE_UTSV_PATH, **kwargs)
+
         self.assertEqual(len(result), 3)
+        self.assertEqual(
+            result[1][0].as_py(), decimal.Decimal("322236.66")
+        )  # rounding decimal
+        self.assertEqual(result[1][1].as_py(), decimal.Decimal("32.33"))  # not rounded
         self.assertEqual(len(result.column_names), 2)
         result_schema = result.schema
-        for index, field in enumerate(result_schema):
-            self.assertEqual(field.name, schema.field(index).name)
+        self.assertEqual(result_schema.field(0).type, "string")
+        self.assertEqual(result_schema.field(1).type, pa.decimal128(20, 2))
 
-        self.assertEqual(result.schema.field(0).type, "string")
+    def test_read_csv_when_decimal_scale_overflows_and_negative_scale(self):
+        schema = pa.schema(
+            [("is_active", pa.string()), ("decimal_value", pa.decimal128(20, -2))]
+        )
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        _add_column_kwargs(
+            ContentType.UNESCAPED_TSV.value,
+            ["is_active", "decimal_value"],
+            ["is_active", "decimal_value"],
+            kwargs,
+        )
 
-    def test_s3_file_to_table_gzip_compressed_sanity(self):
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
+
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
+
+        result = pyarrow_read_csv(OVERFLOWING_DECIMAL_SCALE_UTSV_PATH, **kwargs)
+
+        self.assertEqual(len(result), 3)
+        self.assertEqual(
+            result[1][0].as_py(),
+            decimal.Decimal("322200"),  # consequence of negative scale
+        )  # rounding decimal
+        self.assertEqual(result[1][1].as_py(), decimal.Decimal("00"))
+        self.assertEqual(len(result.column_names), 2)
+        result_schema = result.schema
+        self.assertEqual(result_schema.field(0).type, "string")
+        self.assertEqual(result_schema.field(1).type, pa.decimal128(20, -2))
 
+    def test_read_csv_when_decimal_scale_overflows_with_decimal256(self):
         schema = pa.schema(
-            [("is_active", pa.string()), ("ship_datetime_utc", pa.timestamp("us"))]
+            [("is_active", pa.string()), ("decimal_value", pa.decimal256(20, 2))]
+        )
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        _add_column_kwargs(
+            ContentType.UNESCAPED_TSV.value,
+            ["is_active", "decimal_value"],
+            ["is_active", "decimal_value"],
+            kwargs,
         )
 
-        result = s3_file_to_table(
-            GZIP_COMPRESSED_FILE_UTSV_PATH,
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
+
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
+
+        result = pyarrow_read_csv(OVERFLOWING_DECIMAL_SCALE_UTSV_PATH, **kwargs)
+
+        self.assertEqual(len(result), 3)
+        self.assertEqual(
+            result[1][0].as_py(), decimal.Decimal("322236.66")
+        )  # rounding decimal
+        self.assertEqual(result[1][1].as_py(), decimal.Decimal("32.33"))  # not rounded
+        self.assertEqual(len(result.column_names), 2)
+        result_schema = result.schema
+        self.assertEqual(result_schema.field(0).type, "string")
+        self.assertEqual(result_schema.field(1).type, pa.decimal256(20, 2))
+
+    def test_read_csv_when_decimal_scale_overflows_with_decimal256_and_raise_on_overflow(
+        self,
+    ):
+        schema = pa.schema(
+            [("is_active", pa.string()), ("decimal_value", pa.decimal256(20, 2))]
+        )
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        _add_column_kwargs(
             ContentType.UNESCAPED_TSV.value,
-            ContentEncoding.GZIP.value,
-            ["is_active", "ship_datetime_utc"],
-            None,
-            pa_read_func_kwargs_provider=ReadKwargsProviderPyArrowSchemaOverride(
-                schema=schema
+            ["is_active", "decimal_value"],
+            ["is_active", "decimal_value"],
+            kwargs,
+        )
+
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
+
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
+
+        self.assertRaises(
+            pa.lib.ArrowNotImplementedError,
+            lambda: pyarrow_read_csv(
+                OVERFLOWING_DECIMAL_SCALE_UTSV_PATH,
+                **{**kwargs, RAISE_ON_DECIMAL_OVERFLOW: True},
             ),
         )
 
-        self.assertEqual(len(result), 3)
+    def test_read_csv_when_decimal_scale_overflows_without_any_schema_then_infers(self):
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=None)
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
+
+        result = pyarrow_read_csv(OVERFLOWING_DECIMAL_SCALE_UTSV_PATH, **kwargs)
+
+        # The default behavior of pyarrow is to invalid skip rows
+        self.assertEqual(len(result), 2)
+        self.assertEqual(result[1][0].as_py(), 32.33)  # rounding decimal
+        self.assertEqual(result[1][1].as_py(), 0.4)  # not rounded
         self.assertEqual(len(result.column_names), 2)
         result_schema = result.schema
-        for index, field in enumerate(result_schema):
-            self.assertEqual(field.name, schema.field(index).name)
+        self.assertEqual(result_schema.field(0).type, "string")
+        self.assertEqual(result_schema.field(1).type, pa.float64())
 
-        self.assertEqual(result.schema.field(0).type, "string")
+    def test_read_csv_when_decimal_scale_and_precision_overflow_and_raise_on_overflow(
+        self,
+    ):
+        schema = pa.schema(
+            [("is_active", pa.string()), ("decimal_value", pa.decimal128(5, 2))]
+        )
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        _add_column_kwargs(
+            ContentType.UNESCAPED_TSV.value,
+            ["is_active", "decimal_value"],
+            ["is_active", "decimal_value"],
+            kwargs,
+        )
+
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
+
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
 
-    def test_s3_file_to_table_bz2_compressed_sanity(self):
+        self.assertRaises(
+            pa.lib.ArrowInvalid,
+            lambda: pyarrow_read_csv(OVERFLOWING_DECIMAL_SCALE_UTSV_PATH, **kwargs),
+        )
 
+    def test_read_csv_when_decimal_scale_overflow_and_file_like_obj_passed(self):
         schema = pa.schema(
-            [("is_active", pa.string()), ("ship_datetime_utc", pa.timestamp("us"))]
+            [("is_active", pa.string()), ("decimal_value", pa.decimal128(15, 2))]
+        )
+        kwargs = content_type_to_reader_kwargs(ContentType.UNESCAPED_TSV.value)
+        _add_column_kwargs(
+            ContentType.UNESCAPED_TSV.value,
+            ["is_active", "decimal_value"],
+            ["is_active", "decimal_value"],
+            kwargs,
+        )
+
+        read_kwargs_provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
+
+        kwargs = read_kwargs_provider(ContentType.UNESCAPED_TSV.value, kwargs)
+
+        with open(OVERFLOWING_DECIMAL_SCALE_UTSV_PATH, "rb") as file:
+            result = pyarrow_read_csv(file, **kwargs)
+
+            self.assertEqual(len(result), 3)
+            self.assertEqual(
+                result[1][0].as_py(), decimal.Decimal("322236.66")
+            )  # rounding decimal
+            self.assertEqual(
+                result[1][1].as_py(), decimal.Decimal("32.33")
+            )  # not rounded
+            self.assertEqual(len(result.column_names), 2)
+            result_schema = result.schema
+            self.assertEqual(result_schema.field(0).type, "string")
+            self.assertEqual(result_schema.field(1).type, pa.decimal128(15, 2))
+
+
+class TestWriters(TestCase):
+    def setUp(self):
+        self.table = pa.table({"col1": ["a,b\tc|d", "e,f\tg|h"], "col2": [1, 2]})
+        self.fs = fsspec.filesystem("file")
+        self.base_path = tempfile.mkdtemp()
+        self.fs.makedirs(self.base_path, exist_ok=True)
+
+    def tearDown(self):
+        self.fs.rm(self.base_path, recursive=True)
+
+    def test_write_feather(self):
+        path = f"{self.base_path}/test.feather"
+
+        table_to_file(
+            self.table,
+            path,
+            self.fs,
+            lambda x: path,
+            content_type=ContentType.FEATHER.value,
+        )
+        assert self.fs.exists(path), "file was not written"
+
+        # Verify content
+        result = paf.read_table(path)
+        assert result.equals(self.table)
+
+    def test_write_csv(self):
+        path = f"{self.base_path}/test.csv.gz"
+
+        table_to_file(
+            self.table,
+            path,
+            self.fs,
+            lambda x: path,
+            content_type=ContentType.CSV.value,
+        )
+        assert self.fs.exists(path), "file was not written"
+
+        # Verify content (should be GZIP compressed)
+        with self.fs.open(path, "rb") as f:
+            with gzip.GzipFile(fileobj=f) as gz:
+                content = gz.read().decode("utf-8")
+                # Should be quoted due to commas in data
+                assert '"a,b\tc|d",1' in content
+                assert '"e,f\tg|h",2' in content
+
+    def test_write_tsv(self):
+        path = f"{self.base_path}/test.tsv.gz"
+
+        table_to_file(
+            self.table,
+            path,
+            self.fs,
+            lambda x: path,
+            content_type=ContentType.TSV.value,
+        )
+        assert self.fs.exists(path), "file was not written"
+
+        # Verify content (should be GZIP compressed)
+        with self.fs.open(path, "rb") as f:
+            with gzip.GzipFile(fileobj=f) as gz:
+                content = gz.read().decode("utf-8")
+                # Should be quoted due to tabs in data
+                assert '"a,b\tc|d"\t1' in content
+                assert '"e,f\tg|h"\t2' in content
+
+    def test_write_psv(self):
+        path = f"{self.base_path}/test.psv.gz"
+
+        table_to_file(
+            self.table,
+            path,
+            self.fs,
+            lambda x: path,
+            content_type=ContentType.PSV.value,
+        )
+        assert self.fs.exists(path), "file was not written"
+
+        # Verify content (should be GZIP compressed)
+        with self.fs.open(path, "rb") as f:
+            with gzip.GzipFile(fileobj=f) as gz:
+                content = gz.read().decode("utf-8")
+                # Should be quoted due to pipes in data
+                assert '"a,b\tc|d"|1' in content
+                assert '"e,f\tg|h"|2' in content
+
+    def test_write_unescaped_tsv(self):
+        # Create table without delimiters for unescaped TSV
+        table = pa.table({"col1": ["abc", "def"], "col2": [1, 2]})
+        path = f"{self.base_path}/test.tsv.gz"
+
+        table_to_file(
+            table,
+            path,
+            self.fs,
+            lambda x: path,
+            content_type=ContentType.UNESCAPED_TSV.value,
+        )
+        assert self.fs.exists(path), "file was not written"
+
+        # Verify content (should be GZIP compressed)
+        with self.fs.open(path, "rb") as f:
+            with gzip.GzipFile(fileobj=f) as gz:
+                content = gz.read().decode("utf-8")
+                # With quoting_style="none", strings should not be quoted
+                assert "abc\t1" in content
+                assert "def\t2" in content
+
+    def test_write_orc(self):
+        path = f"{self.base_path}/test.orc"
+
+        table_to_file(
+            self.table,
+            path,
+            self.fs,
+            lambda x: path,
+            content_type=ContentType.ORC.value,
+        )
+        assert self.fs.exists(path), "file was not written"
+
+        # Verify content
+        result = paorc.read_table(path)
+        assert result.equals(self.table)
+
+    def test_write_parquet(self):
+        path = f"{self.base_path}/test.parquet"
+
+        table_to_file(
+            self.table,
+            path,
+            self.fs,
+            lambda x: path,
+            content_type=ContentType.PARQUET.value,
+        )
+        assert self.fs.exists(path), "file was not written"
+
+        # Verify content
+        result = papq.read_table(path)
+        assert result.equals(self.table)
+
+    def test_write_json(self):
+        path = f"{self.base_path}/test.json.gz"
+
+        table_to_file(
+            self.table,
+            path,
+            self.fs,
+            lambda x: path,
+            content_type=ContentType.JSON.value,
+        )
+        assert self.fs.exists(path), "file was not written"
+
+        # Verify content (should be GZIP compressed)
+        with self.fs.open(path, "rb") as f:
+            with gzip.GzipFile(fileobj=f) as gz:
+                content = gz.read().decode("utf-8")
+                # Each line should be a valid JSON object
+                lines = [
+                    line for line in content.split("\n") if line
+                ]  # Skip empty lines
+                assert len(lines) == 2  # 2 records
+                assert json.loads(lines[0]) == {"col1": "a,b\tc|d", "col2": 1}
+                assert json.loads(lines[1]) == {"col1": "e,f\tg|h", "col2": 2}
+
+    def test_write_avro(self):
+        import polars as pl
+
+        path = f"{self.base_path}/test.avro"
+
+        table_to_file(
+            self.table,
+            path,
+            self.fs,
+            lambda x: path,
+            content_type=ContentType.AVRO.value,
+        )
+        assert self.fs.exists(path), "file was not written"
+
+        # Verify content by reading with polars
+        result = pl.read_avro(path).to_arrow()
+        # Cast the result to match the original table's schema
+        # (the round-trip from arrow->polars->arrow casts string to large string)
+        result = result.cast(self.table.schema)
+        assert result.equals(self.table)
+
+
+class TestPyArrowReaders(TestCase):
+    def setUp(self):
+        # Create test data files for reading
+        self.fs = fsspec.filesystem("file")
+        self.base_path = tempfile.mkdtemp()
+        self.fs.makedirs(self.base_path, exist_ok=True)
+
+        # Create test Table
+        self.table = pa.Table.from_pylist(
+            [
+                {"col1": "a,b\tc|d", "col2": 1, "col3": 1.1},
+                {"col1": "e,f\tg|h", "col2": 2, "col3": 2.2},
+                {"col1": "test", "col2": 3, "col3": 3.3},
+            ]
+        )
+
+        # Write test files in different formats
+        self._create_test_files()
+
+    def tearDown(self):
+        self.fs.rm(self.base_path, recursive=True)
+
+    def _create_test_files(self):
+        # Create CSV file (GZIP compressed)
+        csv_path = f"{self.base_path}/test.csv"
+        with self.fs.open(csv_path, "wb") as f:
+            with gzip.GzipFile(fileobj=f, mode="wb") as gz:
+                content = '"a,b\tc|d",1,1.1\n"e,f\tg|h",2,2.2\ntest,3,3.3\n'
+                gz.write(content.encode("utf-8"))
+
+        # Create TSV file (GZIP compressed)
+        tsv_path = f"{self.base_path}/test.tsv"
+        with self.fs.open(tsv_path, "wb") as f:
+            with gzip.GzipFile(fileobj=f, mode="wb") as gz:
+                content = '"a,b\tc|d"\t1\t1.1\n"e,f\tg|h"\t2\t2.2\ntest\t3\t3.3\n'
+                gz.write(content.encode("utf-8"))
+
+        # Create PSV file (GZIP compressed)
+        psv_path = f"{self.base_path}/test.psv"
+        with self.fs.open(psv_path, "wb") as f:
+            with gzip.GzipFile(fileobj=f, mode="wb") as gz:
+                content = '"a,b\tc|d"|1|1.1\n"e,f\tg|h"|2|2.2\ntest|3|3.3\n'
+                gz.write(content.encode("utf-8"))
+
+        # Create unescaped TSV file (GZIP compressed)
+        unescaped_tsv_path = f"{self.base_path}/test_unescaped.tsv"
+        pa.Table.from_pylist(
+            [
+                {"col1": "abc", "col2": 1, "col3": 1.1},
+                {"col1": "def", "col2": 2, "col3": 2.2},
+                {"col1": "ghi", "col2": 3, "col3": 3.3},
+            ]
+        )
+        with self.fs.open(unescaped_tsv_path, "wb") as f:
+            with gzip.GzipFile(fileobj=f, mode="wb") as gz:
+                content = "abc\t1\t1.1\ndef\t2\t2.2\nghi\t3\t3.3\n"
+                gz.write(content.encode("utf-8"))
+
+        # Create Parquet file
+        parquet_path = f"{self.base_path}/test.parquet"
+        with self.fs.open(parquet_path, "wb") as f:
+            papq.write_table(self.table, f)
+
+        # Create Feather file
+        feather_path = f"{self.base_path}/test.feather"
+        with self.fs.open(feather_path, "wb") as f:
+            paf.write_feather(self.table, f)
+
+        # Create JSON file (GZIP compressed)
+        json_path = f"{self.base_path}/test.json"
+        with self.fs.open(json_path, "wb") as f:
+            with gzip.GzipFile(fileobj=f, mode="wb") as gz:
+                # Create NDJSON format - one JSON object per line
+                lines = []
+                for row in self.table.to_pylist():
+                    lines.append(json.dumps(row))
+                content = "\n".join(lines) + "\n"
+                gz.write(content.encode("utf-8"))
+
+        # Create Avro file using polars (since pyarrow delegates to polars for Avro)
+        avro_path = f"{self.base_path}/test.avro"
+        import polars as pl
+
+        pl_df = pl.from_arrow(self.table)
+        pl_df.write_avro(avro_path)
+
+        # Create ORC file
+        orc_path = f"{self.base_path}/test.orc"
+        with self.fs.open(orc_path, "wb") as f:
+            paorc.write_table(self.table, f)
+
+    def test_content_type_to_reader_kwargs(self):
+        # Test CSV kwargs
+        csv_kwargs = content_type_to_reader_kwargs(ContentType.CSV.value)
+        expected_csv = {"parse_options": pacsv.ParseOptions(delimiter=",")}
+        assert (
+            csv_kwargs["parse_options"].delimiter
+            == expected_csv["parse_options"].delimiter
+        )
+
+        # Test TSV kwargs
+        tsv_kwargs = content_type_to_reader_kwargs(ContentType.TSV.value)
+        expected_tsv = {"parse_options": pacsv.ParseOptions(delimiter="\t")}
+        assert (
+            tsv_kwargs["parse_options"].delimiter
+            == expected_tsv["parse_options"].delimiter
+        )
+
+        # Test PSV kwargs
+        psv_kwargs = content_type_to_reader_kwargs(ContentType.PSV.value)
+        expected_psv = {"parse_options": pacsv.ParseOptions(delimiter="|")}
+        assert (
+            psv_kwargs["parse_options"].delimiter
+            == expected_psv["parse_options"].delimiter
+        )
+
+        # Test unescaped TSV kwargs
+        unescaped_kwargs = content_type_to_reader_kwargs(
+            ContentType.UNESCAPED_TSV.value
+        )
+        assert unescaped_kwargs["parse_options"].delimiter == "\t"
+        assert unescaped_kwargs["parse_options"].quote_char is False
+        assert unescaped_kwargs["convert_options"].null_values == [""]
+
+        # Test Parquet kwargs (should be empty)
+        parquet_kwargs = content_type_to_reader_kwargs(ContentType.PARQUET.value)
+        assert parquet_kwargs == {}
+
+        # Test ORC kwargs (should be empty)
+        orc_kwargs = content_type_to_reader_kwargs(ContentType.ORC.value)
+        assert orc_kwargs == {}
+
+        # Test Avro kwargs (should be empty)
+        avro_kwargs = content_type_to_reader_kwargs(ContentType.AVRO.value)
+        assert avro_kwargs == {}
+
+    def test_add_column_kwargs(self):
+        kwargs = {}
+        column_names = ["col1", "col2", "col3"]
+        include_columns = ["col1", "col2"]
+
+        # Test CSV column kwargs
+        _add_column_kwargs(ContentType.CSV.value, column_names, include_columns, kwargs)
+        assert kwargs["read_options"].column_names == column_names
+        assert kwargs["convert_options"].include_columns == include_columns
+
+        # Test Parquet column kwargs
+        kwargs = {}
+        _add_column_kwargs(
+            ContentType.PARQUET.value, column_names, include_columns, kwargs
+        )
+        assert kwargs["columns"] == include_columns
+
+    def test_file_to_table_csv(self):
+        # Test reading CSV with file_to_table
+        csv_path = f"{self.base_path}/test.csv"
+
+        result = file_to_table(
+            csv_path,
+            ContentType.CSV.value,
+            ContentEncoding.GZIP.value,
+            filesystem=self.fs,
+            column_names=["col1", "col2", "col3"],
+        )
+
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        assert result.column("col1").to_pylist() == ["a,b\tc|d", "e,f\tg|h", "test"]
+
+    def test_file_to_table_tsv(self):
+        # Test reading TSV with file_to_table
+        tsv_path = f"{self.base_path}/test.tsv"
+
+        result = file_to_table(
+            tsv_path,
+            ContentType.TSV.value,
+            ContentEncoding.GZIP.value,
+            filesystem=self.fs,
+            column_names=["col1", "col2", "col3"],
+        )
+
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        assert result.column("col1").to_pylist() == ["a,b\tc|d", "e,f\tg|h", "test"]
+
+    def test_file_to_table_psv(self):
+        # Test reading PSV with file_to_table
+        psv_path = f"{self.base_path}/test.psv"
+
+        result = file_to_table(
+            psv_path,
+            ContentType.PSV.value,
+            ContentEncoding.GZIP.value,
+            filesystem=self.fs,
+            column_names=["col1", "col2", "col3"],
         )
 
-        result = s3_file_to_table(
-            BZ2_COMPRESSED_FILE_UTSV_PATH,
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        assert result.column("col1").to_pylist() == ["a,b\tc|d", "e,f\tg|h", "test"]
+
+    def test_file_to_table_unescaped_tsv(self):
+        # Test reading unescaped TSV with file_to_table
+        unescaped_tsv_path = f"{self.base_path}/test_unescaped.tsv"
+
+        result = file_to_table(
+            unescaped_tsv_path,
             ContentType.UNESCAPED_TSV.value,
+            ContentEncoding.GZIP.value,
+            filesystem=self.fs,
+            column_names=["col1", "col2", "col3"],
+        )
+
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        assert result.column("col1").to_pylist() == ["abc", "def", "ghi"]
+
+    def test_file_to_table_parquet(self):
+        # Test reading Parquet with file_to_table
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        result = file_to_table(
+            parquet_path, ContentType.PARQUET.value, filesystem=self.fs
+        )
+
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        assert result.equals(self.table)
+
+    def test_file_to_table_feather(self):
+        # Test reading Feather with file_to_table
+        feather_path = f"{self.base_path}/test.feather"
+
+        result = file_to_table(
+            feather_path, ContentType.FEATHER.value, filesystem=self.fs
+        )
+
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        assert result.equals(self.table)
+
+    def test_file_to_table_json(self):
+        # Test reading JSON with file_to_table
+        json_path = f"{self.base_path}/test.json"
+
+        result = file_to_table(
+            json_path,
+            ContentType.JSON.value,
+            ContentEncoding.GZIP.value,
+            filesystem=self.fs,
+        )
+
+        assert len(result) == 3
+        assert set(result.column_names) == {"col1", "col2", "col3"}
+        assert result.column("col1").to_pylist() == ["a,b\tc|d", "e,f\tg|h", "test"]
+
+    def test_file_to_table_avro(self):
+        # Test reading Avro with file_to_table
+        avro_path = f"{self.base_path}/test.avro"
+
+        result = file_to_table(avro_path, ContentType.AVRO.value, filesystem=self.fs)
+
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        # Avro may have different dtypes, so compare values
+        assert result.column("col1").to_pylist() == ["a,b\tc|d", "e,f\tg|h", "test"]
+
+    def test_file_to_table_orc(self):
+        # Test reading ORC with file_to_table
+        orc_path = f"{self.base_path}/test.orc"
+
+        result = file_to_table(orc_path, ContentType.ORC.value, filesystem=self.fs)
+
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        assert result.equals(self.table)
+
+    def test_file_to_table_with_column_selection(self):
+        # Test reading with column selection
+        csv_path = f"{self.base_path}/test.csv"
+
+        result = file_to_table(
+            csv_path,
+            ContentType.CSV.value,
+            ContentEncoding.GZIP.value,
+            filesystem=self.fs,
+            column_names=["col1", "col2", "col3"],
+            include_columns=["col1", "col2"],
+        )
+
+        assert len(result) == 3
+        assert len(result.column_names) == 2  # Should only have 2 columns
+        assert result.column_names == ["col1", "col2"]
+
+    def test_file_to_table_with_kwargs_provider(self):
+        # Test reading with kwargs provider
+        csv_path = f"{self.base_path}/test.csv"
+        provider = ReadKwargsProviderPyArrowCsvPureUtf8(
+            include_columns=["col1", "col2", "col3"]
+        )
+
+        result = file_to_table(
+            csv_path,
+            ContentType.CSV.value,
+            ContentEncoding.GZIP.value,
+            filesystem=self.fs,
+            column_names=["col1", "col2", "col3"],
+            pa_read_func_kwargs_provider=provider,
+        )
+
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        # With string types provider, all columns should be strings
+        for col_name in result.column_names:
+            assert result.schema.field(col_name).type == pa.string()
+
+    def test_file_to_table_filesystem_inference(self):
+        # Test filesystem inference when no filesystem is provided
+        # Use JSON file since it should work well with inference
+        json_path = f"{self.base_path}/test.json"
+
+        result = file_to_table(
+            json_path,
+            ContentType.JSON.value,
+            ContentEncoding.GZIP.value
+            # No filesystem provided - should be inferred
+        )
+
+        assert len(result) == 3
+        assert set(result.column_names) == {"col1", "col2", "col3"}
+        assert result.column("col1").to_pylist() == ["a,b\tc|d", "e,f\tg|h", "test"]
+
+    def test_file_to_table_unsupported_content_type(self):
+        # Test error handling for unsupported content type
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        with self.assertRaises(NotImplementedError) as context:
+            file_to_table(parquet_path, "unsupported/content-type", filesystem=self.fs)
+
+        assert "not implemented" in str(context.exception)
+
+    def test_file_to_table_bzip2_compression(self):
+        # Test BZIP2 compression handling
+        import bz2
+
+        # Create a BZIP2 compressed CSV file
+        csv_content = '"a,b\tc|d",1,1.1\n"e,f\tg|h",2,2.2\ntest,3,3.3\n'
+        compressed_content = bz2.compress(csv_content.encode("utf-8"))
+
+        bz2_path = f"{self.base_path}/test.csv.bz2"
+        with self.fs.open(bz2_path, "wb") as f:
+            f.write(compressed_content)
+
+        result = file_to_table(
+            bz2_path,
+            ContentType.CSV.value,
             ContentEncoding.BZIP2.value,
-            ["is_active", "ship_datetime_utc"],
-            None,
-            pa_read_func_kwargs_provider=ReadKwargsProviderPyArrowSchemaOverride(
-                schema=schema
-            ),
+            filesystem=self.fs,
+            column_names=["col1", "col2", "col3"],
         )
 
-        self.assertEqual(len(result), 3)
-        self.assertEqual(len(result.column_names), 2)
-        result_schema = result.schema
-        for index, field in enumerate(result_schema):
-            self.assertEqual(field.name, schema.field(index).name)
+        assert len(result) == 3
+        assert result.column_names == ["col1", "col2", "col3"]
+        assert result.column("col1").to_pylist() == ["a,b\tc|d", "e,f\tg|h", "test"]
 
-        self.assertEqual(result.schema.field(0).type, "string")
 
-    def test_s3_file_to_table_when_parquet_sanity(self):
+class TestFileToParquet(TestCase):
+    def setUp(self):
+        # Create test data files for reading
+        self.fs = fsspec.filesystem("file")
+        self.base_path = tempfile.mkdtemp()
+        self.fs.makedirs(self.base_path, exist_ok=True)
+
+        # Create test Table
+        self.table = pa.Table.from_pylist(
+            [
+                {"col1": "a,b\tc|d", "col2": 1, "col3": 1.1},
+                {"col1": "e,f\tg|h", "col2": 2, "col3": 2.2},
+                {"col1": "test", "col2": 3, "col3": 3.3},
+            ]
+        )
+
+        # Write test parquet files
+        self._create_test_files()
 
-        pa_kwargs_provider = lambda content_type, kwargs: {
-            "reader_type": "pyarrow",
-            **kwargs,
-        }
+    def tearDown(self):
+        self.fs.rm(self.base_path, recursive=True)
 
-        result = s3_file_to_table(
-            PARQUET_FILE_PATH,
-            ContentType.PARQUET.value,
-            ContentEncoding.IDENTITY.value,
-            ["n_legs", "animal"],
-            ["n_legs"],
-            pa_read_func_kwargs_provider=pa_kwargs_provider,
+    def _create_test_files(self):
+        # Create basic Parquet file
+        parquet_path = f"{self.base_path}/test.parquet"
+        with self.fs.open(parquet_path, "wb") as f:
+            papq.write_table(self.table, f)
+
+        # Create larger Parquet file with multiple row groups
+        large_table = pa.Table.from_pylist(
+            [{"col1": f"row_{i}", "col2": i, "col3": float(i)} for i in range(1000)]
         )
+        large_parquet_path = f"{self.base_path}/test_large.parquet"
+        with self.fs.open(large_parquet_path, "wb") as f:
+            papq.write_table(
+                large_table, f, row_group_size=100
+            )  # Create multiple row groups
 
-        self.assertEqual(len(result), 6)
-        self.assertEqual(len(result.column_names), 1)
-        schema = result.schema
-        schema_index = schema.get_field_index("n_legs")
-        self.assertEqual(schema.field(schema_index).type, "int64")
+    def test_file_to_parquet_basic(self):
+        # Test basic parquet file reading
+        parquet_path = f"{self.base_path}/test.parquet"
 
-    def test_s3_file_to_table_when_parquet_schema_overridden(self):
+        result = file_to_parquet(parquet_path, filesystem=self.fs)
+
+        assert isinstance(result, papq.ParquetFile)
+        assert result.num_row_groups > 0
+        assert result.metadata.num_rows == 3
+        assert result.metadata.num_columns == 3
+
+        # Verify we can read the data
+        table = result.read()
+        assert len(table) == 3
+        assert table.column_names == ["col1", "col2", "col3"]
+
+    def test_file_to_parquet_with_schema_provider(self):
+        # Test with schema override provider
+        parquet_path = f"{self.base_path}/test.parquet"
 
         schema = pa.schema(
-            [pa.field("animal", pa.string()), pa.field("n_legs", pa.string())]
+            [
+                pa.field("col1", pa.string()),
+                pa.field("col2", pa.string()),  # Override to string
+                pa.field("col3", pa.string()),  # Override to string
+            ]
         )
 
-        pa_kwargs_provider = lambda content_type, kwargs: {
-            "schema": schema,
-            "reader_type": "pyarrow",
-            **kwargs,
-        }
+        provider = ReadKwargsProviderPyArrowSchemaOverride(schema=schema)
 
-        result = s3_file_to_table(
-            PARQUET_FILE_PATH,
+        result = file_to_parquet(
+            parquet_path, filesystem=self.fs, pa_read_func_kwargs_provider=provider
+        )
+
+        assert isinstance(result, papq.ParquetFile)
+        # Note: schema override might not affect ParquetFile metadata,
+        # but should work when reading the table
+        table = result.read()
+        assert len(table) == 3
+
+    def test_file_to_parquet_with_custom_kwargs(self):
+        # Test with custom ParquetFile kwargs
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        result = file_to_parquet(
+            parquet_path,
+            filesystem=self.fs,
+            validate_schema=True,  # Custom kwarg for ParquetFile
+            memory_map=True,  # Another custom kwarg
+        )
+
+        assert isinstance(result, papq.ParquetFile)
+        assert result.metadata.num_rows == 3
+
+    def test_file_to_parquet_filesystem_inference(self):
+        # Test filesystem inference when no filesystem is provided
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        result = file_to_parquet(
+            parquet_path
+            # No filesystem provided - should be inferred
+        )
+
+        assert isinstance(result, papq.ParquetFile)
+        assert result.metadata.num_rows == 3
+        assert result.metadata.num_columns == 3
+
+    def test_file_to_parquet_large_file(self):
+        # Test with larger parquet file (multiple row groups)
+        large_parquet_path = f"{self.base_path}/test_large.parquet"
+
+        result = file_to_parquet(large_parquet_path, filesystem=self.fs)
+
+        assert isinstance(result, papq.ParquetFile)
+        assert result.metadata.num_rows == 1000
+        assert result.num_row_groups > 1  # Should have multiple row groups
+
+        # Test reading specific row groups
+        first_row_group = result.read_row_group(0)
+        assert len(first_row_group) <= 100  # Based on row_group_size=100
+
+    def test_file_to_parquet_metadata_access(self):
+        # Test accessing various metadata properties
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        result = file_to_parquet(parquet_path, filesystem=self.fs)
+
+        # Test metadata access
+        metadata = result.metadata
+        assert metadata.num_rows == 3
+        assert metadata.num_columns == 3
+        assert metadata.num_row_groups >= 1
+
+        # Test schema access
+        schema = result.schema
+        assert len(schema) == 3
+        assert "col1" in schema.names
+        assert "col2" in schema.names
+        assert "col3" in schema.names
+
+        # Test schema_arrow property
+        schema_arrow = result.schema_arrow
+        assert isinstance(schema_arrow, pa.Schema)
+        assert len(schema_arrow) == 3
+
+    def test_file_to_parquet_column_selection(self):
+        # Test reading specific columns
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        result = file_to_parquet(parquet_path, filesystem=self.fs)
+
+        # Read only specific columns
+        table = result.read(columns=["col1", "col2"])
+        assert len(table.column_names) == 2
+        assert table.column_names == ["col1", "col2"]
+        assert len(table) == 3
+
+    def test_file_to_parquet_invalid_content_type(self):
+        # Test error handling for invalid content type
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        with self.assertRaises(ContentTypeValidationError) as context:
+            file_to_parquet(
+                parquet_path,
+                content_type=ContentType.CSV.value,  # Invalid content type
+                filesystem=self.fs,
+            )
+
+        assert "cannot be read into pyarrow.parquet.ParquetFile" in str(
+            context.exception
+        )
+
+    def test_file_to_parquet_invalid_content_encoding(self):
+        # Test error handling for invalid content encoding
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        with self.assertRaises(ContentTypeValidationError) as context:
+            file_to_parquet(
+                parquet_path,
+                content_encoding=ContentEncoding.GZIP.value,  # Invalid encoding
+                filesystem=self.fs,
+            )
+
+        assert "cannot be read into pyarrow.parquet.ParquetFile" in str(
+            context.exception
+        )
+
+    def test_file_to_parquet_different_filesystems(self):
+        # Test with different filesystem implementations
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        # Test with fsspec filesystem
+        result_fsspec = file_to_parquet(parquet_path, filesystem=self.fs)
+        assert isinstance(result_fsspec, papq.ParquetFile)
+        assert result_fsspec.metadata.num_rows == 3
+
+        # Test with None filesystem (inferred)
+        result_inferred = file_to_parquet(parquet_path, filesystem=None)
+        assert isinstance(result_inferred, papq.ParquetFile)
+        assert result_inferred.metadata.num_rows == 3
+
+    def test_file_to_parquet_lazy_loading(self):
+        # Test that ParquetFile provides lazy loading capabilities
+        large_parquet_path = f"{self.base_path}/test_large.parquet"
+
+        result = file_to_parquet(large_parquet_path, filesystem=self.fs)
+
+        # ParquetFile should be created without loading all data
+        assert isinstance(result, papq.ParquetFile)
+        assert result.metadata.num_rows == 1000
+
+        # Test reading only specific columns (lazy loading)
+        partial_table = result.read(columns=["col1", "col2"])
+        assert len(partial_table) == 1000  # All rows but only 2 columns
+        assert partial_table.column_names == ["col1", "col2"]
+
+        # Test reading specific row group (lazy loading)
+        row_group_table = result.read_row_group(0)
+        assert len(row_group_table) <= 100  # Based on row_group_size
+
+    def test_file_to_parquet_performance_timing(self):
+        # Test that performance timing is logged (basic functionality test)
+        parquet_path = f"{self.base_path}/test.parquet"
+
+        # This should complete without error and log timing
+        result = file_to_parquet(parquet_path, filesystem=self.fs)
+
+        assert isinstance(result, papq.ParquetFile)
+        assert result.metadata.num_rows == 3
+
+
+class TestFileToTableFilesystems(TestCase):
+    """Test file_to_table with different filesystem implementations across all content types."""
+
+    def setUp(self):
+        self.tmpdir = tempfile.mkdtemp()
+        self._create_test_files()
+
+    def tearDown(self):
+        import shutil
+
+        shutil.rmtree(self.tmpdir)
+
+    def _create_test_files(self):
+        """Create test files for all supported content types."""
+        # Test data
+        test_data = pa.table(
+            {
+                "id": [1, 2, 3, 4, 5],
+                "name": ["Alice", "Bob", "Charlie", "Diana", "Eve"],
+                "age": [25, 30, 35, 28, 32],
+                "score": [85.5, 92.0, 78.5, 88.0, 95.5],
+            }
+        )
+
+        # File paths
+        self.csv_file = f"{self.tmpdir}/test.csv"
+        self.tsv_file = f"{self.tmpdir}/test.tsv"
+        self.psv_file = f"{self.tmpdir}/test.psv"
+        self.unescaped_tsv_file = f"{self.tmpdir}/test_unescaped.tsv"
+        self.parquet_file = f"{self.tmpdir}/test.parquet"
+        self.feather_file = f"{self.tmpdir}/test.feather"
+        self.json_file = f"{self.tmpdir}/test.json"
+        self.orc_file = f"{self.tmpdir}/test.orc"
+        self.avro_file = f"{self.tmpdir}/test.avro"
+
+        # Create CSV file
+        pacsv.write_csv(
+            test_data,
+            self.csv_file,
+            write_options=pacsv.WriteOptions(delimiter=",", include_header=False),
+        )
+
+        # Create TSV file
+        pacsv.write_csv(
+            test_data,
+            self.tsv_file,
+            write_options=pacsv.WriteOptions(delimiter="\t", include_header=False),
+        )
+
+        # Create PSV file
+        pacsv.write_csv(
+            test_data,
+            self.psv_file,
+            write_options=pacsv.WriteOptions(delimiter="|", include_header=False),
+        )
+
+        # Create unescaped TSV file
+        pacsv.write_csv(
+            test_data,
+            self.unescaped_tsv_file,
+            write_options=pacsv.WriteOptions(
+                delimiter="\t", include_header=False, quoting_style="none"
+            ),
+        )
+
+        # Create Parquet file
+        papq.write_table(test_data, self.parquet_file)
+
+        # Create Feather file
+        paf.write_feather(test_data, self.feather_file)
+
+        # Create JSON file (write as JSONL format)
+        df = test_data.to_pandas()
+        with open(self.json_file, "w") as f:
+            for _, row in df.iterrows():
+                json.dump(row.to_dict(), f)
+                f.write("\n")
+
+        # Create ORC file
+        paorc.write_table(test_data, self.orc_file)
+
+        # Create Avro file
+        try:
+            import polars as pl
+
+            pl_df = pl.from_arrow(test_data)
+            pl_df.write_avro(self.avro_file)
+        except ImportError:
+            # Skip Avro file creation if polars is not available
+            self.avro_file = None
+
+    def _get_filesystems(self, file_path):
+        """Get different filesystem implementations for testing."""
+        # fsspec AbstractFileSystem
+        fsspec_fs = fsspec.filesystem("file")
+
+        # PyArrow filesystem
+        import pyarrow.fs as pafs
+
+        pyarrow_fs = pafs.LocalFileSystem()
+
+        # None for automatic inference
+        auto_infer_fs = None
+
+        return [
+            ("fsspec", fsspec_fs),
+            ("pyarrow", pyarrow_fs),
+            ("auto_infer", auto_infer_fs),
+        ]
+
+    def _assert_table_content(self, table, content_type):
+        """Assert that the loaded table has expected content."""
+        self.assertEqual(len(table), 5, f"Expected 5 rows for {content_type}")
+        self.assertEqual(
+            len(table.columns), 4, f"Expected 4 columns for {content_type}"
+        )
+
+        # Check column names exist (order might vary for some formats)
+        column_names = set(table.column_names)
+        expected_columns = {"id", "name", "age", "score"}
+        self.assertEqual(
+            column_names, expected_columns, f"Column names mismatch for {content_type}"
+        )
+
+    def test_csv_all_filesystems(self):
+        """Test CSV reading with all filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.csv_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.csv_file,
+                    ContentType.CSV.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                    column_names=["id", "name", "age", "score"],
+                )
+                self._assert_table_content(table, f"CSV with {fs_name}")
+
+    def test_tsv_all_filesystems(self):
+        """Test TSV reading with all filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.tsv_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.tsv_file,
+                    ContentType.TSV.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                    column_names=["id", "name", "age", "score"],
+                )
+                self._assert_table_content(table, f"TSV with {fs_name}")
+
+    def test_psv_all_filesystems(self):
+        """Test PSV reading with all filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.psv_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.psv_file,
+                    ContentType.PSV.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                    column_names=["id", "name", "age", "score"],
+                )
+                self._assert_table_content(table, f"PSV with {fs_name}")
+
+    def test_unescaped_tsv_all_filesystems(self):
+        """Test unescaped TSV reading with all filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.unescaped_tsv_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.unescaped_tsv_file,
+                    ContentType.UNESCAPED_TSV.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                    column_names=["id", "name", "age", "score"],
+                )
+                self._assert_table_content(table, f"UNESCAPED_TSV with {fs_name}")
+
+    def test_parquet_all_filesystems(self):
+        """Test Parquet reading with all filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.parquet_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.parquet_file,
+                    ContentType.PARQUET.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                )
+                self._assert_table_content(table, f"PARQUET with {fs_name}")
+
+    def test_feather_all_filesystems(self):
+        """Test Feather reading with all filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.feather_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.feather_file,
+                    ContentType.FEATHER.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                )
+                self._assert_table_content(table, f"FEATHER with {fs_name}")
+
+    def test_json_all_filesystems(self):
+        """Test JSON reading with all filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.json_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.json_file,
+                    ContentType.JSON.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                )
+                self._assert_table_content(table, f"JSON with {fs_name}")
+
+    def test_orc_all_filesystems(self):
+        """Test ORC reading with all filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.orc_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.orc_file,
+                    ContentType.ORC.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                )
+                self._assert_table_content(table, f"ORC with {fs_name}")
+
+    def test_avro_all_filesystems(self):
+        """Test Avro reading with all filesystem types."""
+        if self.avro_file is None:
+            self.skipTest("Avro file creation skipped (polars not available)")
+
+        for fs_name, filesystem in self._get_filesystems(self.avro_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.avro_file,
+                    ContentType.AVRO.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                )
+                self._assert_table_content(table, f"AVRO with {fs_name}")
+
+    def test_column_selection_all_filesystems(self):
+        """Test column selection works with all filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.parquet_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.parquet_file,
+                    ContentType.PARQUET.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                    include_columns=["name", "age"],
+                )
+                self.assertEqual(
+                    len(table.columns), 2, f"Expected 2 columns with {fs_name}"
+                )
+                self.assertEqual(
+                    set(table.column_names),
+                    {"name", "age"},
+                    f"Column selection failed with {fs_name}",
+                )
+
+    def test_kwargs_provider_all_filesystems(self):
+        """Test that kwargs providers work with all filesystem types."""
+
+        def schema_provider(content_type, kwargs):
+            if content_type == ContentType.CSV.value:
+                # Force all columns to be strings
+                kwargs["convert_options"] = pacsv.ConvertOptions(
+                    column_types={
+                        "id": pa.string(),
+                        "name": pa.string(),
+                        "age": pa.string(),
+                        "score": pa.string(),
+                    }
+                )
+            return kwargs
+
+        for fs_name, filesystem in self._get_filesystems(self.csv_file):
+            with self.subTest(filesystem=fs_name):
+                table = file_to_table(
+                    self.csv_file,
+                    ContentType.CSV.value,
+                    ContentEncoding.IDENTITY.value,
+                    filesystem=filesystem,
+                    column_names=["id", "name", "age", "score"],
+                    pa_read_func_kwargs_provider=schema_provider,
+                )
+                # Check that all columns are strings
+                for field in table.schema:
+                    self.assertEqual(
+                        field.type,
+                        pa.string(),
+                        f"Column {field.name} should be string with {fs_name}",
+                    )
+
+    def test_filesystem_auto_inference_consistency(self):
+        """Test that auto-inferred filesystem produces same results as explicit filesystems."""
+        # Use Parquet as it's most reliable across filesystem types
+
+        # Read with auto-inference
+        auto_table = file_to_table(
+            self.parquet_file,
             ContentType.PARQUET.value,
             ContentEncoding.IDENTITY.value,
-            ["n_legs", "animal"],
-            pa_read_func_kwargs_provider=pa_kwargs_provider,
+            filesystem=None,  # Auto-infer
         )
 
-        self.assertEqual(len(result), 6)
-        self.assertEqual(len(result.column_names), 2)
+        # Read with explicit fsspec filesystem
+        fsspec_fs = fsspec.filesystem("file")
+        fsspec_table = file_to_table(
+            self.parquet_file,
+            ContentType.PARQUET.value,
+            ContentEncoding.IDENTITY.value,
+            filesystem=fsspec_fs,
+        )
 
-        result_schema = result.schema
-        for index, field in enumerate(result_schema):
-            self.assertEqual(field.name, schema.field(index).name)
+        # Read with explicit PyArrow filesystem
+        import pyarrow.fs as pafs
+
+        pyarrow_fs = pafs.LocalFileSystem()
+        pyarrow_table = file_to_table(
+            self.parquet_file,
+            ContentType.PARQUET.value,
+            ContentEncoding.IDENTITY.value,
+            filesystem=pyarrow_fs,
+        )
+
+        # All should produce equivalent results
+        self.assertTrue(
+            auto_table.equals(fsspec_table),
+            "Auto-inferred result should match fsspec result",
+        )
+        self.assertTrue(
+            auto_table.equals(pyarrow_table),
+            "Auto-inferred result should match PyArrow result",
+        )
 
-        self.assertEqual(result.schema.field(1).type, "string")
+    def test_error_handling_all_filesystems(self):
+        """Test error handling works consistently across filesystem types."""
+        for fs_name, filesystem in self._get_filesystems(self.parquet_file):
+            with self.subTest(filesystem=fs_name):
+                # Test unsupported content type
+                with self.assertRaises(NotImplementedError):
+                    file_to_table(
+                        self.parquet_file,
+                        "UNSUPPORTED_TYPE",
+                        ContentEncoding.IDENTITY.value,
+                        filesystem=filesystem,
+                    )
+
+                # Test non-existent file
+                with self.assertRaises((FileNotFoundError, OSError)):
+                    file_to_table(
+                        f"{self.tmpdir}/non_existent.parquet",
+                        ContentType.PARQUET.value,
+                        ContentEncoding.IDENTITY.value,
+                        filesystem=filesystem,
+                    )