[FEAT] Enable MicroPartitions by default

benchmarking/parquet/conftest.py

@@ -41,7 +41,7 @@ def boto3_get_object_read(path: str, columns: list[str] | None = None) -> pa.Tab
 
 
 def daft_native_read(path: str, columns: list[str] | None = None) -> pa.Table:
-    tbl = daft.table.Table.read_parquet(path, columns=columns)
+    tbl = daft.table.MicroPartition.read_parquet(path, columns=columns)
     return tbl.to_arrow()
 
 
@@ -76,7 +76,7 @@ def fn(files: list[str]) -> list[pa.Table]:
 
 
 def daft_bulk_read(paths: list[str], columns: list[str] | None = None) -> list[pa.Table]:
-    tables = daft.table.LegacyTable.read_parquet_bulk(paths, columns=columns)
+    tables = daft.table.Table.read_parquet_bulk(paths, columns=columns)
     return [t.to_arrow() for t in tables]
 
 

daft/dataframe/dataframe.py

@@ -33,7 +33,7 @@
 from daft.logical.builder import LogicalPlanBuilder
 from daft.runners.partitioning import PartitionCacheEntry, PartitionSet
 from daft.runners.pyrunner import LocalPartitionSet
-from daft.table import Table
+from daft.table import MicroPartition
 from daft.viz import DataFrameDisplay
 
 if TYPE_CHECKING:
@@ -184,7 +184,7 @@ def __iter__(self) -> Iterator[Dict[str, Any]]:
                     yield row
 
     @DataframePublicAPI
-    def iter_partitions(self) -> Iterator[Union[Table, "RayObjectRef"]]:
+    def iter_partitions(self) -> Iterator[Union[MicroPartition, "RayObjectRef"]]:
         """Begin executing this dataframe and return an iterator over the partitions.
 
         Each partition will be returned as a daft.Table object (if using Python runner backend)
@@ -236,27 +236,27 @@ def _from_pydict(cls, data: Dict[str, InputListType]) -> "DataFrame":
                 f"Expected all columns to be of the same length, but received columns with lengths: {column_lengths}"
             )
 
-        data_vpartition = Table.from_pydict(data)
+        data_vpartition = MicroPartition.from_pydict(data)
         return cls._from_tables(data_vpartition)
 
     @classmethod
     def _from_arrow(cls, data: Union["pa.Table", List["pa.Table"]]) -> "DataFrame":
         """Creates a DataFrame from a pyarrow Table."""
         if not isinstance(data, list):
             data = [data]
-        data_vpartitions = [Table.from_arrow(table) for table in data]
+        data_vpartitions = [MicroPartition.from_arrow(table) for table in data]
         return cls._from_tables(*data_vpartitions)
 
     @classmethod
     def _from_pandas(cls, data: Union["pd.DataFrame", List["pd.DataFrame"]]) -> "DataFrame":
         """Creates a Daft DataFrame from a pandas DataFrame."""
         if not isinstance(data, list):
             data = [data]
-        data_vpartitions = [Table.from_pandas(df) for df in data]
+        data_vpartitions = [MicroPartition.from_pandas(df) for df in data]
         return cls._from_tables(*data_vpartitions)
 
     @classmethod
-    def _from_tables(cls, *parts: Table) -> "DataFrame":
+    def _from_tables(cls, *parts: MicroPartition) -> "DataFrame":
         """Creates a Daft DataFrame from a single Table.
 
         Args:
@@ -1036,7 +1036,7 @@ def _construct_show_display(self, n: int) -> "DataFrameDisplay":
                 if seen >= n:
                     break
 
-            preview_partition = Table.concat(tables)
+            preview_partition = MicroPartition.concat(tables)
             if len(preview_partition) > n:
                 preview_partition = preview_partition.slice(0, n)
             elif len(preview_partition) < n:

daft/dataframe/preview.py

@@ -2,12 +2,12 @@
 
 from dataclasses import dataclass
 
-from daft.table import Table
+from daft.table import MicroPartition
 
 
 @dataclass(frozen=True)
 class DataFramePreview:
     """A class containing all the metadata/data required to preview a dataframe."""
 
-    preview_partition: Table | None
+    preview_partition: MicroPartition | None
     dataframe_num_rows: int | None