Question

import numpy as np
from numpy import genfromtxt

train = genfromtxt('/Users/hhimanshu/Downloads/dataset/digitrecognizer/a.csv', delimiter=',', names=True)

当我运行形状信息时，我得到了

# if train is ndarray
print 'shape(Tuple of array dimensions) = ', train.shape
print 'dimension(Number of array dimensions) = ', train.ndim
print 'size(Number of elements in array) = ', train.size
# print 'data type = ', train.dtype
train['label']
train['pixel783']

我得到了

shape(Tuple of array dimensions) =  (9,)
dimension(Number of array dimensions) =  1
size(Number of elements in array) =  9
array([ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.])

数据看起来像

label,pixel0,pixel1,pixel2,pixel3,pixel4,pixel5,pixel6,pixel7,pixel8,pixel9,pixel10,pixel11,pixel12,pixel13,pixel14,pixel15,pixel16,pixel17,pixel18,pixel19,pixel20,pixel21,pixel22,pixel23,pixel24,pixel25,pixel26,pixel27,pixel28,pixel29,pixel30,pixel31,pixel32,pixel33,pixel34,pixel35,pixel36,pixel37,pixel38,pixel39,pixel40,pixel41,pixel42,pixel43,pixel44,pixel45,pixel46,pixel47,pixel48,pixel49,pixel50,pixel51,pixel52,pixel53,pixel54,pixel55,pixel56,pixel57,pixel58,pixel59,pixel60,pixel61,pixel62,pixel63,pixel64,pixel65,pixel66,pixel67,pixel68,pixel69,pixel70,pixel71,pixel72,pixel73,pixel74,pixel75,pixel76,pixel77,pixel78,pixel79,pixel80,pixel81,pixel82,pixel83,pixel84,pixel85,pixel86,pixel87,pixel88,pixel89,pixel90,pixel91,pixel92,pixel93,pixel94,pixel95,pixel96,pixel97,pixel98,pixel99,pixel100,pixel101,pixel102,pixel103,pixel104,pixel105,pixel106,pixel107,pixel108,pixel109,pixel110,pixel111,pixel112,pixel113,pixel114,pixel115,pixel116,pixel117,pixel118,pixel119,pixel120,pixel121,pixel122,pixel123,pixel124,pixel125,pixel126,pixel127,pixel128,pixel129,pixel130,pixel131,pixel132,pixel133,pixel134,pixel135,pixel136,pixel137,pixel138,pixel139,pixel140,pixel141,pixel142,pixel143,pixel144,pixel145,pixel146,pixel147,pixel148,pixel149,pixel150,pixel151,pixel152,pixel153,pixel154,pixel155,pixel156,pixel157,pixel158,pixel159,pixel160,pixel161,pixel162,pixel163,pixel164,pixel165,pixel166,pixel167,pixel168,pixel169,pixel170,pixel171,pixel172,pixel173,pixel174,pixel175,pixel176,pixel177,pixel178,pixel179,pixel180,pixel181,pixel182,pixel183,pixel184,pixel185,pixel186,pixel187,pixel188,pixel189,pixel190,pixel191,pixel192,pixel193,pixel194,pixel195,pixel196,pixel197,pixel198,pixel199,pixel200,pixel201,pixel202,pixel203,pixel204,pixel205,pixel206,pixel207,pixel208,pixel209,pixel210,pixel211,pixel212,pixel213,pixel214,pixel215,pixel216,pixel217,pixel218,pixel219,pixel220,pixel221,pixel222,pixel223,pixel224,pixel225,pixel226,pixel227,pixel228,pixel229,pixel230,pixel231,pixel232,pixel233,pixel234,pixel235,pixel236,pixel237,pixel238,pixel239,pixel240,pixel241,pixel242,pixel243,pixel244,pixel245,pixel246,pixel247,pixel248,pixel249,pixel250,pixel251,pixel252,pixel253,pixel254,pixel255,pixel256,pixel257,pixel258,pixel259,pixel260,pixel261,pixel262,pixel263,pixel264,pixel265,pixel266,pixel267,pixel268,pixel269,pixel270,pixel271,pixel272,pixel273,pixel274,pixel275,pixel276,pixel277,pixel278,pixel279,pixel280,pixel281,pixel282,pixel283,pixel284,pixel285,pixel286,pixel287,pixel288,pixel289,pixel290,pixel291,pixel292,pixel293,pixel294,pixel295,pixel296,pixel297,pixel298,pixel299,pixel300,pixel301,pixel302,pixel303,pixel304,pixel305,pixel306,pixel307,pixel308,pixel309,pixel310,pixel311,pixel312,pixel313,pixel314,pixel315,pixel316,pixel317,pixel318,pixel319,pixel320,pixel321,pixel322,pixel323,pixel324,pixel325,pixel326,pixel327,pixel328,pixel329,pixel330,pixel331,pixel332,pixel333,pixel334,pixel335,pixel336,pixel337,pixel338,pixel339,pixel340,pixel341,pixel342,pixel343,pixel344,pixel345,pixel346,pixel347,pixel348,pixel349,pixel350,pixel351,pixel352,pixel353,pixel354,pixel355,pixel356,pixel357,pixel358,pixel359,pixel360,pixel361,pixel362,pixel363,pixel364,pixel365,pixel366,pixel367,pixel368,pixel369,pixel370,pixel371,pixel372,pixel373,pixel374,pixel375,pixel376,pixel377,pixel378,pixel379,pixel380,pixel381,pixel382,pixel383,pixel384,pixel385,pixel386,pixel387,pixel388,pixel389,pixel390,pixel391,pixel392,pixel393,pixel394,pixel395,pixel396,pixel397,pixel398,pixel399,pixel400,pixel401,pixel402,pixel403,pixel404,pixel405,pixel406,pixel407,pixel408,pixel409,pixel410,pixel411,pixel412,pixel413,pixel414,pixel415,pixel416,pixel417,pixel418,pixel419,pixel420,pixel421,pixel422,pixel423,pixel424,pixel425,pixel426,pixel427,pixel428,pixel429,pixel430,pixel431,pixel432,pixel433,pixel434,pixel435,pixel436,pixel437,pixel438,pixel439,pixel440,pixel441,pixel442,pixel443,pixel444,pixel445,pixel446,pixel447,pixel448,pixel449,pixel450,pixel451,pixel452,pixel453,pixel454,pixel455,pixel456,pixel457,pixel458,pixel459,pixel460,pixel461,pixel462,pixel463,pixel464,pixel465,pixel466,pixel467,pixel468,pixel469,pixel470,pixel471,pixel472,pixel473,pixel474,pixel475,pixel476,pixel477,pixel478,pixel479,pixel480,pixel481,pixel482,pixel483,pixel484,pixel485,pixel486,pixel487,pixel488,pixel489,pixel490,pixel491,pixel492,pixel493,pixel494,pixel495,pixel496,pixel497,pixel498,pixel499,pixel500,pixel501,pixel502,pixel503,pixel504,pixel505,pixel506,pixel507,pixel508,pixel509,pixel510,pixel511,pixel512,pixel513,pixel514,pixel515,pixel516,pixel517,pixel518,pixel519,pixel520,pixel521,pixel522,pixel523,pixel524,pixel525,pixel526,pixel527,pixel528,pixel529,pixel530,pixel531,pixel532,pixel533,pixel534,pixel535,pixel536,pixel537,pixel538,pixel539,pixel540,pixel541,pixel542,pixel543,pixel544,pixel545,pixel546,pixel547,pixel548,pixel549,pixel550,pixel551,pixel552,pixel553,pixel554,pixel555,pixel556,pixel557,pixel558,pixel559,pixel560,pixel561,pixel562,pixel563,pixel564,pixel565,pixel566,pixel567,pixel568,pixel569,pixel570,pixel571,pixel572,pixel573,pixel574,pixel575,pixel576,pixel577,pixel578,pixel579,pixel580,pixel581,pixel582,pixel583,pixel584,pixel585,pixel586,pixel587,pixel588,pixel589,pixel590,pixel591,pixel592,pixel593,pixel594,pixel595,pixel596,pixel597,pixel598,pixel599,pixel600,pixel601,pixel602,pixel603,pixel604,pixel605,pixel606,pixel607,pixel608,pixel609,pixel610,pixel611,pixel612,pixel613,pixel614,pixel615,pixel616,pixel617,pixel618,pixel619,pixel620,pixel621,pixel622,pixel623,pixel624,pixel625,pixel626,pixel627,pixel628,pixel629,pixel630,pixel631,pixel632,pixel633,pixel634,pixel635,pixel636,pixel637,pixel638,pixel639,pixel640,pixel641,pixel642,pixel643,pixel644,pixel645,pixel646,pixel647,pixel648,pixel649,pixel650,pixel651,pixel652,pixel653,pixel654,pixel655,pixel656,pixel657,pixel658,pixel659,pixel660,pixel661,pixel662,pixel663,pixel664,pixel665,pixel666,pixel667,pixel668,pixel669,pixel670,pixel671,pixel672,pixel673,pixel674,pixel675,pixel676,pixel677,pixel678,pixel679,pixel680,pixel681,pixel682,pixel683,pixel684,pixel685,pixel686,pixel687,pixel688,pixel689,pixel690,pixel691,pixel692,pixel693,pixel694,pixel695,pixel696,pixel697,pixel698,pixel699,pixel700,pixel701,pixel702,pixel703,pixel704,pixel705,pixel706,pixel707,pixel708,pixel709,pixel710,pixel711,pixel712,pixel713,pixel714,pixel715,pixel716,pixel717,pixel718,pixel719,pixel720,pixel721,pixel722,pixel723,pixel724,pixel725,pixel726,pixel727,pixel728,pixel729,pixel730,pixel731,pixel732,pixel733,pixel734,pixel735,pixel736,pixel737,pixel738,pixel739,pixel740,pixel741,pixel742,pixel743,pixel744,pixel745,pixel746,pixel747,pixel748,pixel749,pixel750,pixel751,pixel752,pixel753,pixel754,pixel755,pixel756,pixel757,pixel758,pixel759,pixel760,pixel761,pixel762,pixel763,pixel764,pixel765,pixel766,pixel767,pixel768,pixel769,pixel770,pixel771,pixel772,pixel773,pixel774,pixel775,pixel776,pixel777,pixel778,pixel779,pixel780,pixel781,pixel782,pixel783
1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,188,255,94,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,191,250,253,93,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,123,248,253,167,10,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,80,247,253,208,13,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,29,207,253,235,77,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,54,209,253,253,88,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,93,254,253,238,170,17,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23,210,254,253,159,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,16,209,253,254,240,81,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,27,253,253,254,13,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,20,206,254,254,198,7,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,168,253,253,196,7,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,20,203,253,248,76,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,22,188,253,245,93,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,103,253,253,191,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,89,240,253,195,25,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,15,220,253,253,80,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,94,253,253,253,94,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,89,251,253,250,131,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,214,218,95,0,0,0,0,0,0,0,0,0,0,0,0

我认为形状应该是(9, 784)，不是吗？我在这里做的不是什么？

更新
但是，当我执行以下操作时（删除names=True）

train = genfromtxt(fname='/Users/hhimanshu/Downloads/dataset/digitrecognizer/a.csv', delimiter=',')

并运行形状信息

# if train is ndarray
print 'shape(Tuple of array dimensions) = ', train.shape
print 'dimension(Number of array dimensions) = ', train.ndim
print 'size(Number of elements in array) = ', train.size
# print 'data type = ', train.dtype
train['label']

我得到正确的尺寸，但标签信息丢失了

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-136-53c55437f109> in <module>()
      4 print 'size(Number of elements in array) = ', train.size
      5 # print 'data type = ', train.dtype
----> 6 train['label']
      7 # train['pixel1']

ValueError: field named label not found.

shape(Tuple of array dimensions) =  (10, 785)
dimension(Number of array dimensions) =  2
size(Number of elements in array) =  7850

Answer 1

当您使用names = True时，这是一种特殊情况：genfromtxt返回包含连续数据的astructured数组（由@'Pierre GM'指出）。

我希望你单独加载标题和数据（例如带有跳数）。

Answer 2

在numpy中，ndarray的所有元素必须具有相同的类型。这种类型可以很简单（例如float或int）或更复杂，因为它是几种子类型的组合（例如[('name', '|S16'),('value',float)]。在这种情况下，我们谈论的是structured array也就是说，一个带有结构dtype的ndarray。注意你可以定义一个结构化的dtype，其中所有的子元素都有相同的类型，比如[('first',float),('second',float)]。通常，结构化数组的一个元素被调用记录和每个子元素称为字段。

例如考虑数组：

example = np.array([(1,2,3),(4,5,6)], dtype=[('A',int),('B',int),('C',int)])

example是一个由两个元素（(1,2,3)和(4,5,6)）组成的结构化数组，每个元素（或“记录”）都有3个字段。因为它只有2个元素，所以它的形状为(2,)。您可以使用len(example.dtype.names)获取字段数。这个数组的大小是2，因为我们只有两个元素。

因为它是一个结构化数组，我们可以单独访问每个字段：

>>> example['A']
array([1, 4])

总之：

如documentation of np.genfromtxt中所述，当您使用names参数时，您特别要求np.genfromtxt返回结构化数组。

输出的每个元素对应于文件的有效行，并且具有与文件中的列一样多的子元素（“字段”）（784）。在这里，由于您有9个有效行，因此输出的形状为(9,)。

因为你没有给出dtype，genfromtxt假设每个字段都是float，所以输出的dtype由下式给出：

np.dtype([(name, float) for name in list_of_names])

从您的文件的第一个有效行中读取list_of_names。

如果您现在不使用names=True，实际上是告诉np.genfromtxt将您的文件读取为浮动的2D数组，因此(9,784)形状。但是因为在这种情况下，输出不再是结构化数组，您不能通过其名称访问每个列（因为它没有任何）。

请注意，在这种特殊情况下（所有字段都是浮点数），您可以使用

从结构化视图切换到非结构化视图

train.reshape((-1,1)).view(float)

Answer 3

这应该解决它，阅读标题，然后是其他，正如georgesl所建议的那样。我通过文件句柄传递所有内容来简化它：

fh=open(myfile,'r')
header=fh.readline().rstrip("\n").split(",")
data=np.genfromtxt(fh,delimiter=",")
fh.close()

Numpy：获取不正确的形状信息

3 个答案: