[SPARK-12300] [SQL] [PYSPARK] fix schema inferance on local collections

Current schema inference for local python collections halts as soon as there are no NullTypes. This is different than when we specify a sampling ratio of 1.0 on a distributed collection. This could result in incomplete schema information.

Author: Holden Karau <holden@us.ibm.com>

Closes #10275 from holdenk/SPARK-12300-fix-schmea-inferance-on-local-collections.

1 files changed, 3 insertions, 7 deletions

diff --git a/python/pyspark/sql/context.py b/python/pyspark/sql/context.py
index b05aa2f5c4..ba6915a123 100644
--- a/python/pyspark/sql/context.py
+++ b/python/pyspark/sql/context.py
@@ -18,6 +18,7 @@
 import sys
 import warnings
 import json
+from functools import reduce
 
 if sys.version >= '3':
     basestring = unicode = str
@@ -236,14 +237,9 @@ class SQLContext(object):
         if type(first) is dict:
             warnings.warn("inferring schema from dict is deprecated,"
                           "please use pyspark.sql.Row instead")
-        schema = _infer_schema(first)
+        schema = reduce(_merge_type, map(_infer_schema, data))
         if _has_nulltype(schema):
-            for r in data:
-                schema = _merge_type(schema, _infer_schema(r))
-                if not _has_nulltype(schema):
-                    break
-            else:
-                raise ValueError("Some of types cannot be determined after inferring")
+            raise ValueError("Some of types cannot be determined after inferring")
         return schema
 
     def _inferSchema(self, rdd, samplingRatio=None):