如何将 numpy 列表输出到 cvs 文件 python 3.6 中

Question

>我有一个python列表如下：

[(numpy.datetime64('2000-04-01'), 'pie'),
(numpy.datetime64('2000-04-01'), 'apple'),
(numpy.datetime64('2000-04-01'), 'orange'),
(numpy.datetime64('2000-04-01'), 'mango'),
(numpy.datetime64('2000-04-01'), 'pears')]

我希望它到csv文件输出如下

|    date    |  item |
+------------+-------+
| 2000-04-01 | pie   |
| 2000-04-01 | apple |
|     ...    |  ...  |

Answer 1

我终于能够输出为 csv 文件。

import csv
with open('date_prop.csv','w',newline='') as out:
csv_out=csv.writer(out)
csv_out.writerow(['date','item'])
for row in f2:
csv_out.writerow(row)

Answer 2

直接的 Python 文件写入：

In [116]: with open('test','w') as f:
...:     print('   date   item', file=f)
...:     for t in alist:
...:         print('%s   %s'%t, file=f)
...:         
In [117]: cat test
date   item
2000-04-01   pie
2000-04-01   apple
2000-04-01   orange
2000-04-01   mango
2000-04-01   pears

关键是str(np.datetime64('2000-04-01')的格式是您想要的。

对象 dtype 数组可以用右fmtsavetxt编写：

In [121]: arr = np.array(alist)
In [122]: arr
Out[122]: 
array([[numpy.datetime64('2000-04-01'), 'pie'],
[numpy.datetime64('2000-04-01'), 'apple'],
[numpy.datetime64('2000-04-01'), 'orange'],
[numpy.datetime64('2000-04-01'), 'mango'],
[numpy.datetime64('2000-04-01'), 'pears']], dtype=object)
In [123]: np.savetxt('test', arr, fmt='%s', delimiter=',', header='date item')
In [124]: cat test
# date item
2000-04-01,pie
2000-04-01,apple
2000-04-01,orange
2000-04-01,mango
2000-04-01,pears

我还可以创建一个结构化数组并使用savetxt编写它

In [125]: arr = np.array(alist, dtype='datetime64[D], U10')
In [126]: arr
Out[126]: 
array([('2000-04-01', 'pie'), ('2000-04-01', 'apple'),
('2000-04-01', 'orange'), ('2000-04-01', 'mango'),
('2000-04-01', 'pears')], dtype=[('f0', '<M8[D]'), ('f1', '<U10')])
In [127]: np.savetxt('test', arr, fmt='%s', delimiter=',  ', header='date   item
...: ')
In [128]: cat test
# date   item
2000-04-01,  pie
2000-04-01,  apple
2000-04-01,  orange
2000-04-01,  mango
2000-04-01,  pears