Github user viirya commented on a diff in the pull request:
https://github.com/apache/spark/pull/18444#discussion_r128174597
--- Diff: python/pyspark/sql/types.py ---
@@ -915,6 +916,91 @@ def _parse_datatype_json_value(json_value):
long: LongType,
})
+# Mapping Python array types to Spark SQL DataType
+# We should be careful here. The size of these types in python depends on C
+# implementation. We need to make sure that this conversion does not lose
any
+# precision. Also, JVM only support signed types, when converting unsigned
types,
+# keep in mind that it required 1 more bit when stored as singed types.
+#
+# Reference for C integer size, see:
+# ISO/IEC 9899:201x specification, chapter 5.2.4.2.1 Sizes of integer
types <limits.h>.
+# Reference for python array typecode, see:
+# https://docs.python.org/2/library/array.html
+# https://docs.python.org/3.6/library/array.html
+# Reference for JVM's supported integral types:
+# http://docs.oracle.com/javase/specs/jvms/se8/html/jvms-2.html#jvms-2.3.1
+
+_array_signed_int_typecode_ctype_mappings = {
+ 'b': ctypes.c_byte,
+ 'h': ctypes.c_short,
+ 'i': ctypes.c_int,
+ 'l': ctypes.c_long,
+}
+
+_array_unsigned_int_typecode_ctype_mappings = {
+ 'B': ctypes.c_ubyte,
+ 'H': ctypes.c_ushort,
+ 'I': ctypes.c_uint,
+ 'L': ctypes.c_ulong
+}
+
+
+def _int_size_to_type(size):
+ """
+ Return the Scala type from the size of integers.
+ """
+ if size <= 8:
+ return ByteType
+ if size <= 16:
+ return ShortType
+ if size <= 32:
+ return IntegerType
+ if size <= 64:
+ return LongType
+
+# The list of all supported array typecodes is stored here
+_array_type_mappings = {
+ # Warning: Actual properties for float and double in C is not
specified in C.
+ # On almost every system supported by both python and JVM, they are
IEEE 754
+ # single-precision binary floating-point format and IEEE 754
double-precision
+ # binary floating-point format. And we do assume the same thing here
for now.
+ 'f': FloatType,
+ 'd': DoubleType
+}
+
+# compute array typecode mappings for signed integer types
+for _typecode in _array_signed_int_typecode_ctype_mappings.keys():
+ size =
ctypes.sizeof(_array_signed_int_typecode_ctype_mappings[_typecode]) * 8
+ dt = _int_size_to_type(size)
+ if dt is not None:
+ _array_type_mappings[_typecode] = dt
+
+# compute array typecode mappings for unsigned integer types
+for _typecode in _array_unsigned_int_typecode_ctype_mappings.keys():
+ # JVM does not have unsigned types, so use signed types that is at
list 1
--- End diff --
nit: typo `at list`.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
