fix create DataFrame from pandas

Author: Davies Liu

python/pyspark/sql/context.py

@@ -351,6 +351,8 @@ def createDataFrame(self, data, schema=None, samplingRatio=None):
         :return: a DataFrame
 
         >>> l = [('Alice', 1)]
+        >>> sqlCtx.createDataFrame(l).collect()
+        [Row(_1=u'Alice', _2=1)]
         >>> sqlCtx.createDataFrame(l, ['name', 'age']).collect()
         [Row(name=u'Alice', age=1)]
 
@@ -359,6 +361,8 @@ def createDataFrame(self, data, schema=None, samplingRatio=None):
         [Row(age=1, name=u'Alice')]
 
         >>> rdd = sc.parallelize(l)
+        >>> sqlCtx.createDataFrame(rdd).collect()
+        [Row(_1=u'Alice', _2=1)]
         >>> df = sqlCtx.createDataFrame(rdd, ['name', 'age'])
         >>> df.collect()
         [Row(name=u'Alice', age=1)]
@@ -377,14 +381,17 @@ def createDataFrame(self, data, schema=None, samplingRatio=None):
         >>> df3 = sqlCtx.createDataFrame(rdd, schema)
         >>> df3.collect()
         [Row(name=u'Alice', age=1)]
+
+        >>> sqlCtx.createDataFrame(df.toPandas()).collect()  # doctest: +SKIP
+        [Row(name=u'Alice', age=1)]
         """
         if isinstance(data, DataFrame):
             raise TypeError("data is already a DataFrame")
 
         if has_pandas and isinstance(data, pandas.DataFrame):
-            data = self._sc.parallelize(data.to_records(index=False))
             if schema is None:
                 schema = list(data.columns)
+            data = [r.tolist() for r in data.to_records(index=False)]
 
         if not isinstance(data, RDD):
             try:
@@ -399,7 +406,8 @@ def createDataFrame(self, data, schema=None, samplingRatio=None):
         if isinstance(schema, (list, tuple)):
             first = data.first()
             if not isinstance(first, (list, tuple)):
-                raise ValueError("each row in `rdd` should be list or tuple")
+                raise ValueError("each row in `rdd` should be list or tuple, "
+                                 "but got %r" % type(first))
             row_cls = Row(*schema)
             schema = self._inferSchema(data.map(lambda r: row_cls(*r)), samplingRatio)
 

python/pyspark/sql/types.py

@@ -637,15 +637,14 @@ def _infer_schema(row):
     if isinstance(row, dict):
         items = sorted(row.items())
 
-    elif isinstance(row, tuple):
+    elif isinstance(row, (tuple, list)):
         if hasattr(row, "_fields"):  # namedtuple
             items = zip(row._fields, tuple(row))
         elif hasattr(row, "__FIELDS__"):  # Row
             items = zip(row.__FIELDS__, tuple(row))
-        elif all(isinstance(x, tuple) and len(x) == 2 for x in row):
-            items = row
         else:
-            raise ValueError("Can't infer schema from tuple")
+            names = ['_%d' % i for i in range(1, len(row) + 1)]
+            items = zip(names, row)
 
     elif hasattr(row, "__dict__"):  # object
         items = sorted(row.__dict__.items())
@@ -812,17 +811,10 @@ def convert_struct(obj):
         if obj is None:
             return
 
-        if isinstance(obj, tuple):
-            if hasattr(obj, "_fields"):
-                d = dict(zip(obj._fields, obj))
-            elif hasattr(obj, "__FIELDS__"):
-                d = dict(zip(obj.__FIELDS__, obj))
-            elif all(isinstance(x, tuple) and len(x) == 2 for x in obj):
-                d = dict(obj)
-            else:
-                raise ValueError("unexpected tuple: %s" % str(obj))
+        if isinstance(obj, (tuple, list)):
+            return tuple(conv(v) for v, conv in zip(obj, converters))
 
-        elif isinstance(obj, dict):
+        if isinstance(obj, dict):
             d = obj
         elif hasattr(obj, "__dict__"):  # object
             d = obj.__dict__