writing an array to h5 in python

Question

Using the h5py module I'm trying to simply read in data from one h5 file, do some basic calculations with the data and write it back into a new h5 file. All is well, except when trying to write the dataset.

so far I have:

f = h5py.File(inData,'r')
dset = f['/DATA/DATA/']
HH = dset[...,0]

HHdB = (10*numpy.log10(HH*HH)) - 83

outfile = h5py.File(outData, 'w')
f.create_dataset('/DATA/DATA/', data=(HHdB))

This returns me the error: "ValueError: unable to create dataset (Dataset: Unable to initialize object)", which I don't understand.

Im a newbie so any help would be much appreciated!

unutbu · Accepted Answer · 2014-10-22 10:51:16Z

1

f.create_dataset should be outfile.create_dataset, since f is the File opened in read mode, while outfile is a File opened in write mode.

By the way, if you use the h5py.Files as a context manager in a with-statement, the file will automatically be closed for you (and written to disk) when Python leaves the with-statement.

import numpy
import h5py

with h5py.File(inData,'r') as f:
    dset = f['/DATA/DATA/']

HH = dset[...,0]
HHdB = (10*numpy.log10(HH*HH)) - 83

with h5py.File(outData, 'w') as outfile:
    outfile.create_dataset('/DATA/DATA/', data=HHdB)

edited Oct 22, 2014 at 10:51

answered Oct 22, 2014 at 10:41

unutbu

886k197 gold badges1.9k silver badges1.7k bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

GeoMonkey Over a year ago

Well thats embarrassing. It was some code I copied from somewhere else and I forgot to change it. Thanks for the help (and for not laughing at me)!

usethedeathstar Over a year ago

was going to mention the with-statement, though on some older machines (at work or so), this might not yet work (with-statement was introduced in python 2.5)

Collectives™ on Stack Overflow

writing an array to h5 in python

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related