CODEWITHSUNDEEP
pythoncsvnumpypandas
Ohm
Ohm

Reputation: 2442

Loading csv of different columns number with pandas

I have a csv file in which there are always two first columns, but with varying number of columns for different files. The csv can look like this:

Gondi,4012,227,233,157,158,149,158
Gondi,4013,227,231,156,159,145,153
Gondu,4014,228,233,157,158,145,153
Gondu,4015,227,231,156,159,149,158

For now I am working with NumPy, and my code for loading this data is:

import numpy as np
def readfile(fname):
    with open(fname) as f:
       ncols = len(f.readline().split(','))
    name = np.loadtxt(fname, delimiter=',', usecols=[0],dtype=str)
    ind  = np.loadtxt(fname, delimiter=',', usecols=[1],dtype=int)
    data = np.loadtxt(fname, delimiter=',', usecols=range(2,ncols),dtype=int)
    return data,name,ind

Can I do the same thing with pandas more efficiently?

Upvotes: 2

Views: 99

Answers (1)

jezrael
jezrael

Reputation: 862501

I think you can use read_csv and iloc for select first, second and other columns:

import pandas as pd
import io

temp=u"""Gondi,4012,227,233,157,158,149,158
Gondi,4013,227,231,156,159,145,153
Gondu,4014,228,233,157,158,145,153
Gondu,4015,227,231,156,159,149,158"""
#after testing replace io.StringIO(temp) to filename
df = pd.read_csv(io.StringIO(temp), header=None)
print df

name = df.iloc[:,0]
print name
0    Gondi
1    Gondi
2    Gondu
3    Gondu
Name: 0, dtype: object

ind = df.iloc[:,1]
print ind
0    4012
1    4013
2    4014
3    4015
Name: 1, dtype: int64

data = df.iloc[:,2:]
print data
     2    3    4    5    6    7
0  227  233  157  158  149  158
1  227  231  156  159  145  153
2  228  233  157  158  145  153
3  227  231  156  159  149  158

Upvotes: 1

Related Questions