You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A good example is the ngmix photometry which has array and matrix columns. Some of these are covariance matrices, which are naturally represented in memory as a 9x9 matrix not 81 different columns.
A simple approach is for easyaccess to internally check that the user has requested an array or matrix column, such as cm_pars_cov, and then expand it in the query. E.g. the user could write
select
cm_pars,
cm_pars_cov,
from
....
and this would internally be expanded to
select
cm_pars_1,
cm_pars_2,
....
cm_pars_cov_1_1,
cm_pars_cov_1_2,
...
from
And then when writing out a fits file or returning a numpy array, the data gets packed into single columns cm_pars and cm_pars_cov.
The text was updated successfully, but these errors were encountered:
A good example is the ngmix photometry which has array and matrix columns. Some of these are covariance matrices, which are naturally represented in memory as a
9x9
matrix not 81 different columns.A simple approach is for
easyaccess
to internally check that the user has requested an array or matrix column, such ascm_pars_cov
, and then expand it in the query. E.g. the user could writeand this would internally be expanded to
And then when writing out a fits file or returning a numpy array, the data gets packed into single columns
cm_pars
andcm_pars_cov
.The text was updated successfully, but these errors were encountered: