NumPy 的数据类型

dt = np.dtype('>i4')
dt.byteorder
# '>'
dt.itemsize
# 4
dt.name
# 'int32'
# 对应的数组标量类型是 int32
dt.type is np.int32
# True
np.float64 == np.dtype(np.float64) == np.dtype('float64')
# True
np.float64 == np.dtype(np.float64).type
# True

dt = np.dtype([('name', np.unicode_, 16), ('grades', np.float64, (2,))])
dt['name']
# dtype('<U16')
dt['grades']
# dtype(('<f8', (2,)))

x = np.array([('Sarah', (8.0, 7.0)), ('John', (6.0, 7.0))], dtype=dt)
x[1]
# ('John', [6., 7.])
x[1]['grades']
# array([6.,  7.])
type(x[1])
# <class 'numpy.void'>
type(x[1]['grades'])
# <class 'numpy.ndarray'>

# 使用数组标量类型
np.dtype(np.int16) # dtype('int16')
# 结构化类型，一个字段名“f1”，包含int16:
np.dtype([('f1', np.int16)]) # dtype([('f1', '<i2')])

# 结构化类型，一个名为“f1”的字段
# 其本身包含一个具有一个字段的结构化类型：
np.dtype([('f1', [('f1', np.int16)])])
# dtype([('f1', [('f1', '<i2')])])

# 结构化类型，两个字段：第一个字段包含无符号int
# 第二个字段包含int 32：
np.dtype([('f1', np.uint64), ('f2', np.int32)])
# dtype([('f1', '<u8'), ('f2', '<i4')])

# 使用数组协议类型字符串：
np.dtype([('a','f8'),('b','S10')])
# dtype([('a', '<f8'), ('b', 'S10')])

# 使用逗号分隔的字段格式。形状为（2,3）
np.dtype("i4, (2,3)f8")
# dtype([('f0', '<i4'), ('f1', '<f8', (2, 3))])

# 使用元组，int 是固定类型，3 表示字段的形状
# void 是一种灵活的类型，这里的大小为10：
np.dtype([('hello',(np.int64,3)),('world',np.void,10)])
# dtype([('hello', '<i8', (3,)), ('world', 'V10')])

# 将 int16 细分为 2 个 int8，称为x和y
# 0 和 1 是以字节为单位的偏移量：
np.dtype((np.int16, {'x':(np.int8,0), 'y':(np.int8,1)}))
# dtype((numpy.int16, [('x', 'i1'), ('y', 'i1')]))

# 使用字典。两个名为“gender”和“age”的字段：
np.dtype({'names':['gender','age'], 'formats':['S1',np.uint8]})
# dtype([('gender', 'S1'), ('age', 'u1')])

# 偏移量（字节），这里是0和25：
np.dtype({'surname':('S25',0),'age':(np.uint8,25)})
# dtype([('surname', 'S25'), ('age', 'u1')])

np.dtype(np.int32)      # 32-bit integer
np.dtype(np.complex128) # 128-bit complex floating-point number
np.dtype(float)   # Python-compatible floating-point number
np.dtype(int)     # Python-compatible integer
np.dtype(object)  # Python object
np.dtype('b')  # byte, native byte order
np.dtype('>H') # big-endian unsigned short
np.dtype('<f') # little-endian single-precision float
np.dtype('d')  # double-precision floating-point number
np.dtype('i4')   # 32-bit signed integer
np.dtype('f8')   # 64-bit floating-point number
np.dtype('c16')  # 128-bit complex floating-point number
np.dtype('a25')  # 25-length zero-terminated bytes
np.dtype('U25')  # 25-character string
np.dtype('uint32')   # 32-bit unsigned integer
np.dtype('float64')  # 64-bit floating-point number
np.dtype((np.void, 10))  # 10-byte wide data block
np.dtype(('U', 10))   # 10-character unicode string
np.dtype((np.int32, (2,2)))          # 2 x 2 integer sub-array
np.dtype(('i4, (2,3)f8, f4', (2,3))) # 2 x 3 structured sub-array
# fields big (big-endian 32-bit integer) and
# little (little-endian 32-bit integer)
np.dtype([('big', '>i4'), ('little', '<i4')])
# fields R, G, B, A, each being an unsigned 8-bit integer
np.dtype([('R','u1'), ('G','u1'), ('B','u1'), ('A','u1')])
# fields r, g, b, a, each being an 8-bit unsigned integer
np.dtype({'names': ['r','g','b','a'],
               'formats': [np.uint8, np.uint8, np.uint8, np.uint8]})
# fields r and b (with the given titles), both being 8-bit
# unsigned integers, the first at byte position 0 from the
# start of the field and the second at position 2
np.dtype({'names': ['r','b'], 'formats': ['u1', 'u1'],
          'offsets': [0, 2],
          'titles': ['Red pixel', 'Blue pixel']})
# field col1 (10-character string at byte position 0),
# col2 (32-bit float at byte position 10),
# and col3 (integers at byte position 14)
np.dtype({'col1': ('U10', 0), 'col2': (np.float32, 10),
		'col3': (int, 14)})
# 32-bit integer, whose first two bytes are interpreted
# as an integer via field real, and the
# following two bytes via field imag.
np.dtype((np.int32,{'real':(np.int16, 0),'imag':(np.int16, 2)}))
# 32-bit integer, which is interpreted as consisting of
# a sub-array of shape (4,) containing 8-bit integers:
np.dtype((np.int32, (np.int8, 4)))
# 32-bit integer, containing fields r, g, b, a that interpret
# the 4 bytes in the integer as four unsigned integers:
np.dtype(('i4', [('r','u1'),('g','u1'),('b','u1'),('a','u1')]))

np.result_type(3, np.arange(7, dtype='i1'))
# dtype('int8')
np.result_type('i4', 'c8')
# dtype('complex128')
np.result_type(3.0, -2)
# dtype('float64')

defaults = {np.dtype('int64'): np.int32,
            np.dtype('float64'): np.float32}
before = 1.
np.array(before, dtype=defaults.get(np.result_type(before), None))