Pfthb

I was useing pysal.weights to calculate the spatial lag variable('price') with Gaussian kernel weight , it succeeded in find all the closest kW.neighbors but some rows did not get kW.weights, thus not all of the rows got the lag variable(price).

None of the rows of the coordinates geometries or price columns are nan or zero. So I am not sure what went wrong.

error are as follows:

/Users/xxx/miniconda3/lib/python3.6/site-packages/libpysal/weights/distance.py:645: RuntimeWarning: invalid value encountered in true_divide
 zi = np.array([dict(list(zip(ni, di)))[nid] for nid in nids]) / bw[i]
/Users/xxx/miniconda3/lib/python3.6/site-packages/libpysal/weights/weights.py:171: UserWarning: The weights matrix is not fully connected. There are 796 components
 warnings.warn("The weights matrix is not fully connected. There are %d components" % self.n_components)
/Users/xxx/miniconda3/lib/python3.6/site-packages/libpysal/weights/weights.py:171: UserWarning: The weights matrix is not fully connected. There are 796 components
 warnings.warn("The weights matrix is not fully connected. There are %d components" % self.n_components)

Here's the code:

lag_vars=['price']
kW = lp.weights.Kernel.from_dataframe(df.loc[:,lag_vars+['geometry']], fixed=False, function='gaussian', k=10)
kW = fill_diagonal(kW, 0)
kW.transform = 'r'
WX = lp.weights.lag_spatial(kW, df.loc[:,lag_vars])
WXtable = pd.DataFrame(WX, columns=['lag_'.format(name) for name in lag_vars])
fd_lag = pd.concat((df,WXtable),axis=1)