chunk(0) shortcut

Author: dougbrn

src/nested_pandas/nestedframe/io.py

@@ -136,8 +136,16 @@ def read_parquet(
             field_names = [table.column_names[i] for i in indices]
 
             # Use iterchunks to process chunks of each column
-            chunked_arrays = [pa.concat_arrays(list(table.column(i).iterchunks())) for i in indices]
-
+            chunked_arrays = []
+            for i in indices:
+                column = table.column(i)
+                if len(column.chunks) == 1:
+                    # If there is only one chunk, use it directly
+                    # avoid copy in concat_arrays
+                    chunked_arrays.append(column.chunk(0))
+                else:
+                    # Otherwise, concatenate all chunks
+                    chunked_arrays.append(pa.concat_arrays(list(column.iterchunks())))
             structs[col] = pa.StructArray.from_arrays(
                 chunked_arrays,  # Child arrays
                 field_names,  # Field names