"column pointers (size n+1) or col indices (size nzmax) start from 0"

[green...@pobox.com]

Hi,

Could you expand a little on this option for csc->p

"column pointers (size n+1) or col indices (size nzmax) start from 0"

Are there examples of both schemes?

How does the solver know which scheme you have supplied as input?

[Bartolomeo Stellato]

Hi,

Matrices are defined in Compressed Sparse Column format (CSC). "p" indicates the column pointers.

Have a look at these links for more details:

• https://people.sc.fsu.edu/~jburkardt/data/cc/cc.html
  
• https://docs.scipy.org/doc/scipy-0.19.0/reference/generated/scipy.sparse.csc_matrix.html

The part "col indices (size nzmax)" is related to the internal sparse triplet format representation that OSQP uses when forming the KKT matrix. You should ignore it.

As a user, you should only provide matrices in CSC format (default from Matlab sparse and Scipy sparse).

Let me know if it helps.

Bartolomeo

--
You received this message because you are subscribed to the Google Groups "OSQP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osqp+unsubscribe@googlegroups.com.
To post to this group, send email to os...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/osqp/0657df2c-2301-4e25-bfd1-5ee1dc612434%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[green...@pobox.com]

It was the second part "col indices (size nzmax)" that was of concern, and I will indeed ignore it, thanks.

I was hoping I was doing something wrong that on being fixed would give better adherence to constraints, but no such luck.