[SPARK-10392] [SQL] Pyspark - Wrong DateType support on JDBC connection

This PR addresses issue [SPARK-10392](https://issues.apache.org/jira/browse/SPARK-10392)
The problem is that for "start of epoch" date (01 Jan 1970) PySpark class DateType returns 0 instead of the `datetime.date` due to implementation of its return statement

Issue reproduction on master:
```
>>> from pyspark.sql.types import *
>>> a = DateType()
>>> a.fromInternal(0)
0
>>> a.fromInternal(1)
datetime.date(1970, 1, 2)
```

Author: 0x0FFF <programmerag@gmail.com>

Closes #8556 from 0x0FFF/SPARK-10392.

python/pyspark/sql/tests.py

@@ -168,6 +168,11 @@ def test_decimal_type(self):
         t3 = DecimalType(8)
         self.assertNotEqual(t2, t3)
 
+    # regression test for SPARK-10392
+    def test_datetype_equal_zero(self):
+        dt = DateType()
+        self.assertEqual(dt.fromInternal(0), datetime.date(1970, 1, 1))
+
 
 class SQLTests(ReusedPySparkTestCase):
 

python/pyspark/sql/types.py

@@ -168,10 +168,12 @@ def needConversion(self):
         return True
 
     def toInternal(self, d):
-        return d and d.toordinal() - self.EPOCH_ORDINAL
+        if d is not None:
+            return d.toordinal() - self.EPOCH_ORDINAL
 
     def fromInternal(self, v):
-        return v and datetime.date.fromordinal(v + self.EPOCH_ORDINAL)
+        if v is not None:
+            return datetime.date.fromordinal(v + self.EPOCH_ORDINAL)
 
 
 class TimestampType(AtomicType):