# Read sparse matrix in Harwell-Boeing Exchange Format

I'm currently creating some really large matrices, and hope they will end up reasonably sparse. I would like to toy with them in sage some day. Looking for a suitable standardized storage format, I just learned about the Harwell-Boeing Exchange Format. **Is there some functionality provided by sage to read such files?**

If not, should I rather read the file and construct a matrix from it using a standard matrix constructor, or should I create my own matrix subclass whith an internal format that closely resembles the file format?

Perhaps you could try the Matrix Market format, it is described in the link you posted. You can [read](http://www.scipy.org/doc/api_docs/SciPy.io.mmio.html) this format using scipy, which ships with sage.