Python-从包含分隔符的文本文件中读取经/纬度坐标

Question

Originally, the text file were for the use of GMT (The Generic Mapping Tools). The format is similar as following:

122   55
122   56
122.5 57
>
123   25.25
123   25.27

where '>' is recognized as a separation sign for different segments of lines.

Now, I'm using Basemap from mpl_toolkits to plot lines on a map. All I need is a 2D numpy array to pass coordinates into a function more or less the same as plt.plot or plt.scatter .

Here's a simple solution which I came up with:

import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.basemap import Basemap
from io import StringIO

file = open("latlon.txt", "r")
data = file.read()

data = data.replace('>','nan\tnan') 
# Use nan,nan to isolate different segments of lines in plt.plot
# Also, np.genfromtxt cannot directly read such kinds of data
#  due to inconsistent of columns for certain lines in the file,
#  hence the replacement
line_xy = np.genfromtxt(StringIO(unicode(data)))

(plotting stuffs...)

I found this way kind of tricky and not really like it... Is there any solution which is more commonly, explicitly or formally used for such case? Any advice is welcome.

Answer 1

If you have access to the Pandas toolkit, then the pandas.read_csv() function can be used to parse your data, and turn it into an array of floats:

import pandas as pd
df = pd.read_csv('data.txt', sep='\s+', header=None)
array = df.apply(pd.to_numeric, args=('coerce',)).values

Here, the read_csv() will treat the columns in your datafile as separated by whitespace, and return a DataFrame of strings. The apply() line then converts those into floating-point numbers, coercing invalid entries (like your '>') into NaNs. The .values attribute then extracts the contents of the DataFrame as an ordinary numpy.ndarray . This gives:

array([[ 122.  ,   55.  ],
       [ 122.  ,   56.  ],
       [ 122.5 ,   57.  ],
       [    nan,     nan],
       [ 123.  ,   25.25],
       [ 123.  ,   25.27]])