Hi,

I'm trying to do a very standard spatial analysis. The aim is to run ML and 2SLS estimations with spregress. I've successfully run this several times and then after I changed one of the excel files I was using it suddenly stopped working, or better, it does work, but it creates empty matrices. Here's my log:

.
. clear all

. import excel "network_contiguity_MLE_FULL.xlsx", sheet("network_contiguity_MLE_FULL") firstrow
(34 vars, 6,633 obs)
. .
. gen y = AIM15 - AIM13
(153 missing values generated)
.
. xtset id
panel variable: id (balanced)

. spset id
Sp dataset
data: cross sectional
spatial-unit id: _ID (equal to id)
coordinates: none
linked shapefile: none

.
. mata

: mata matuse W_network_full_U /*Upload Matrix X and Vector v from file wmatrix*/
(loading X[6650,6650], v[6650,1])

: end
.
. spmatrix spfrommata ZW = X v
(weighting matrix contains 6650 islands)

.mata

:sum(X)
4725

The last line is to show that the mata matrix it is not empty, yet it creates a matrix with as many islands as number of observations, so I guess empty?

This code still works using a different mata matrix, but except for the position of 1s and 0s they look identical to me.
I've been struggling for days with this, and I really cannot figure it out.
Thanks a lot for any help.