fix for dask

dask · Jan 19, 2023 · a888be9 · a888be9
1 parent f3c6f2c
commit a888be9
Showing 1 changed file with 5 additions and 2 deletions.
diff --git a/fastparquet/writer.py b/fastparquet/writer.py
@@ -695,7 +695,10 @@ def write_column(f, data0, selement, compression=None, datapage_version=None,
         algorithm = compression
 
     # output thrift metadata
-    s = parquet_thrift.Statistics(max=max, min=min, null_count=global_num_nulls) if stats else None
+    if stats:
+        s = parquet_thrift.Statistics(max=max, min=min, null_count=global_num_nulls)
+    else:
+        s = parquet_thrift.Statistics(null_count=global_num_nulls)
 
     kvm = []
     if isinstance(name, (list, tuple)):
@@ -770,7 +773,7 @@ def make_row_group(f, data, schema, compression=None, stats=True):
             if isinstance(stats, int):
                 st = stats
             elif stats == "auto":
-                st = coldata.dtype.kind in ["i", "u", "f", "M"] or len(data) < 10_000
+                st = coldata.dtype.kind in ["i", "u", "f", "M"]
             else:
                 st = column.name in stats
             chunk = write_column(f, coldata, column,