numpy-Python-选择性导入.txt文件的一部分

Question

In my data.txt file, there are 2 types of lines.

1. Normal data: 16 numbers separated by spaces with a '\\n' appended at the end.

2. Incomplete data: In the process of writing the data into data.txt, the writing-in of the last line is always interrupted by the STOP command. Thus, it is always incomplete, egit can have 10 numbers and no '\\n'

Two questions:

a. How can I import the whole file EXCEPT the last incomplete line into Python?

I notice that

# Load the .txt file in
myData = np.loadtxt('twenty_z_up.txt')

is quite "strict" in the sense that when the last incomplete line exists there, the file cannot be imported. The imported .txt file has to be a nice matrix.

b. Occasionally, I make timestamps on the first entry of a line for experiment purpose. Say I have my 1st timestamp at the start of line 2, and my 2nd stamp at the start of line 5. How can I import only from line 2 to line 5 into Python?

=============================== Updates: Qa is solved ================================

myData = np.genfromtxt('fast_walking_pocket.txt', skip_footer=1)

will help discard the final incomplete row

Answer 1

You can try pandas which provides a use function read_csv to load the data more easily.

Example data:

a b c d e f g h i j k l m n o p
a b c d e f g h i j k l m n o p
a b c d e f g h i j k l m n o p
a b c d e f g h i j k l m n o p
a b c d e f g h i j k l m n o p
a b c d e f g h i j

For your Q1, you can load the data by:

In [27]: import pandas as pd

In [28]: df = pd.read_csv('test.txt', sep=' ', header=None, skipfooter=1)

DataFrame is a useful structure which can help you to process data easier. To get a numpy array, simply get the values attribute of the DataFrame .

In [33]: df.values
Out[33]: 
array([['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm',
        'n', 'o', 'p'],
       ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm',
        'n', 'o', 'p'],
       ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm',
        'n', 'o', 'p'],
       ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm',
        'n', 'o', 'p'],
       ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm',
        'n', 'o', 'p']], dtype=object)

For your Q2, you can get the second and the fifth line by

In [36]: df.ix[[1, 4]]
Out[36]:
  0  1  2  3  4  5  6  7  8  9  10 11 12 13 14 15
1  a  b  c  d  e  f  g  h  i  j  k  l  m  n  o  p
4  a  b  c  d  e  f  g  h  i  j  k  l  m  n  o  p

Answer 2

To answer your 'b' question.

Assume you have this file (called '/tmp/lines.txt'):

line 1
2013:10:15
line 3
line 4
2010:8:15
line 6 

You can use the linecache module:

>>> import linecache
>>> linecache.getline('/tmp/lines.txt', 2)
'2013:10:15\n'

So you can parse this time directly:

>>> import datetime as dt
>>>dt.datetime.strptime(linecache.getline('/tmp/lines.txt',2).strip(),'%Y:%m:%d')
datetime.datetime(2013, 10, 15, 0, 0)

Edit

Multiple lines:

>>> li=[]
>>> for i in (2,5):
...    li.append(linecache.getline('/tmp/lines.txt', i).strip())
... 
>>> li
['2013:10:15', '2010:8:15']

Or:

>>> lines={}
>>> for i in (2,5):
...    lines[i]=linecache.getline('/tmp/lines.txt', i).strip()
... 
>>> lines
{2: '2013:10:15', 5: '2010:8:15'}

Or a range:

>>> lines={}
>>> for i in range(2,6):
...    lines[i]=linecache.getline('/tmp/lines.txt', i).strip()
... 
>>> lines
{2: '2013:10:15', 3: 'line 3', 4: 'line 4', 5: '2010:8:15'}

Answer 3

Question a:

np.genfromtxt('twenty_z_up.txt',skip_footer=1)

Qustion b:

np.genfromtxt('twenty_z_up.txt',skip_footer=1)[2:5]