[SPARK-21658][SQL][PYSPARK] Add default None for value in na.replace …

…in PySpark

## What changes were proposed in this pull request?
JIRA issue: https://issues.apache.org/jira/browse/SPARK-21658

Add default None for value in `na.replace` since `Dataframe.replace` and `DataframeNaFunctions.replace` are alias.

The default values are the same now.
```
>>> df = sqlContext.createDataFrame([('Alice', 10, 80.0)])
>>> df.replace({"Alice": "a"}).first()
Row(_1=u'a', _2=10, _3=80.0)
>>> df.na.replace({"Alice": "a"}).first()
Row(_1=u'a', _2=10, _3=80.0)
```

## How was this patch tested?
Existing tests.

cc viirya

Author: byakuinss <grace.chinhanyu@gmail.com>

Closes #18895 from byakuinss/SPARK-21658.

python/pyspark/sql/dataframe.py

@@ -1403,6 +1403,16 @@ def replace(self, to_replace, value=None, subset=None):
         |null|  null|null|
         +----+------+----+
 
+        >>> df4.na.replace('Alice').show()
+        +----+------+----+
+        | age|height|name|
+        +----+------+----+
+        |  10|    80|null|
+        |   5|  null| Bob|
+        |null|  null| Tom|
+        |null|  null|null|
+        +----+------+----+
+
         >>> df4.na.replace(['Alice', 'Bob'], ['A', 'B'], 'name').show()
         +----+------+----+
         | age|height|name|
@@ -1837,7 +1847,7 @@ def fill(self, value, subset=None):
 
     fill.__doc__ = DataFrame.fillna.__doc__
 
-    def replace(self, to_replace, value, subset=None):
+    def replace(self, to_replace, value=None, subset=None):
         return self.df.replace(to_replace, value, subset)
 
     replace.__doc__ = DataFrame.replace.__doc__