CODEWITHSUNDEEP
pythondatetimenumpy
Apollo
Apollo

Reputation: 9062

Load Data from CSV using Numpy

Data

HOD,2012-1-3,1,5000
HOD,2012-1-4,1,5000
HOD,2012-1-5,1,5000
HOD,2012-1-6,1,5000
HOD,2012-1-9,1,5000
HOD,2012-1-10,1,5000

myData = np.genfromtxt(inputFile, dtype=[('Symbol',str),('Date', 'datetime64[D]'),('Value', int),('Allocation', long)], delimiter=',')
print myData

Output

('', datetime.datetime(1969, 12, 31, 0, 0), 1, 5000L)
('', datetime.datetime(1969, 12, 31, 0, 0), 1, 5000L)
('', datetime.datetime(1969, 12, 31, 0, 0), 1, 5000L)
('', datetime.datetime(1969, 12, 31, 0, 0), 1, 5000L)
('', datetime.datetime(1969, 12, 31, 0, 0), 1, 5000L)

Why are my strings lost?

Upvotes: 1

Views: 113

Answers (1)

falsetru
falsetru

Reputation: 369244

String dtype require you to specify length:

dtype=[('Symbol', 'S3'), ...]

or

dtype=[('Symbol', (str, 3)), ...]

If you want arbitrary-length string, specify object as a type; it will allow any object to be assigned.

Upvotes: 1

Related Questions