Use to_spark()

databricks · Sep 6, 2019 · f813bb7 · f813bb7
1 parent e9becfc
commit f813bb7
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/databricks/koalas/generic.py b/databricks/koalas/generic.py
@@ -639,12 +639,12 @@ def to_json(self, path=None, compression='uncompressed', num_files=None, **kwarg
         kdf = self
         if isinstance(self, ks.Series):
             kdf = self._kdf
-        sdf = kdf._sdf
+        sdf = kdf.to_spark()
 
         if num_files is not None:
             sdf = sdf.repartition(num_files)
 
-        builder = sdf.select(self._internal.data_columns).write.mode("overwrite")
+        builder = sdf.write.mode("overwrite")
         OptionUtils._set_opts(builder, compression=compression)
         builder.options(**kwargs).format("json").save(path)