I would like to know how to import file with multiple delimiter.
I have the following line:
"1,000";"2,000";"3,000"
How can I import the data with numpy?
I have the following code
data=numpy.loadtxt(filepath,delimiter=';')
how can I pass a 2nd delimiter for the " "
Kind Regards
pandas.read_csv can read such a file. It allows you to control the delimiter and the decimal point character.
Here's my file delim.dat:
"1,000";"2,000";"3,000"
"5,000";"6,000";"7,000"
"8,000";"9,000";"9,100"
"9,250";"9,500";"9,990"
Use the arguments delimiter=';' and decimal=',' in pandas.read_csv:
In [11]: import pandas as pd
In [12]: df = pd.read_csv('delim.dat', sep=';', decimal=',', header=None)
In [13]: df
Out[13]:
0 1 2
0 1.00 2.0 3.00
1 5.00 6.0 7.00
2 8.00 9.0 9.10
3 9.25 9.5 9.99
You can also use numpy.genfromtxt, but you'll have to use the converters argument to convert each field from bytes to floating point. For example,
In [54]: def myconvert(s):
...: return float(s.strip(b'"').replace(b',', b'.'))
...:
...:
In [55]: a = np.genfromtxt('delim.dat', delimiter=';', converters={k: myconvert for k in range(3)})
In [56]: a
Out[56]:
array([[1. , 2. , 3. ],
[5. , 6. , 7. ],
[8. , 9. , 9.1 ],
[9.25, 9.5 , 9.99]])
You have some weird delimiters, not to mention the use of commas in the middle of the numeric literals. Numpy isn't set up to handle any of these things by default, so you'll have to parse your input by hand a little bit before you pass it off to Numpy. You can do so using some regular expressions:
import re
from io import StringIO
# fake file-like object for demonstration
f = StringIO('''"1,000";"2,000";"3,000"''')
s = re.sub('";?"?', ' ', f.read())
s = re.sub(',', '.', s)
arr = np.fromstring(s, sep=' ')
print(arr)
Output:
[1. 2. 3.]
dtype=stris supposed to help. You probably should not do that.