CODEWITHSUNDEEP
pythondatetimecsvnumpy
jtam
jtam

Reputation: 902

Using loadtext for CSV file with date/times

I am new to Python (converting from Matlab) and I am having trouble reading a CSV file. Basically, I want create an array with the sequence of date/times in one variable and the corresponding data values (humidity measurements) in another variable (I want to skip the 2nd column). The file format is as follows:

07-24-14 01:00:01 PM,%RH,36.988
07-24-14 02:00:01 PM,%RH,40.832
...

I am using the numpy loadtxt function as follows (note: there are 21 headerlines in the file):

def datestr2num(s):
    return datetime.strptime(s,'%m-%d-%y %I:%M:%S %p')

dates,vals = np.loadtxt('File.csv',usecols=(0,2),skiprows=21,converters={0:datestr2num},delimiter=',',unpack=True)

I am getting the following error:

TypeError: float() argument must be a string or a number

Thanks for your help in advance!

Upvotes: 1

Views: 1423

Answers (1)

farenorth
farenorth

Reputation: 10781

An alternate approach would be to use a float time representation. I often use matplotlib time stamps:

from matplotlib.dates import date2num

def datestr2num(s):
    return date2num(datetime.strptime(s, '%m-%d-%y %I:%M:%S %p'))

Or you could use the python built-in 'time':

import time

def datestr2num(s):
    return time.mktime(datetime.strptime(s, '%m-%d-%y %I:%M:%S %p').timetuple())

If you'll be using one of these representations for your time anyway, this might work well. If you want datetimes, use dano's solution.

Upvotes: 1

Related Questions