Details
-
Type:
Bug
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: Qserv
-
Labels:
-
Story Points:4
-
Epic Link:
-
Sprint:DB_F19_07
-
Team:Data Access and Database
Description
qserv_areaspec_ellipse has semi-major and semi-minor axis parameters. These were documented to be in arcsec, but the code appears to instantiate a sphgeom::Ellipse assuming the values are in degrees. I have just changed the documentation to correspond to this, but I think that is actually wrong, because the scisql UDF that is added to the WHERE clause requires those parameters to be in arcsec.
So I think the correct fix is to change the documentation back to arcsec and have the sphgeom::Ellipse constructed using arcsec rather than degrees.
Attachments
Issue Links
- relates to
-
- To Do
Activity
I'm not sure what "make consistent ... in all layers above" means: "make above layers consistent with the degree/arcsec split in the scisql_ functions", or "make the circle and ellipse size parameters consistent with each other at every layer except scisql_ "? If it's the latter, then to make sure I've understood correctly, is the proposed answer w.r.t. size parameters:
- scisql_* : circle: degrees, ellipse: arcseconds
- qserv_* : circle: degrees, ellipse: degrees
- ADQL: CIRCLE() function: degrees; REGION() function's quasi-STC-S (if/when supported): degrees for both circles and ellipses
?
Gregory Dubois-Felsmann not quite. The proposed solution is:
- scisql_: circle radius: degrees, ellipse semi-axes: arcseconds (per published docs, and same as it ever was)
- qserv_: exact same as scisql_, now in word and deed
ADQL: degrees everywhere, per spec
(The point of the change being to make scisql_ and qserv_ congruent, and remove some lies in the docs and bugs in the qserv code, while leaving scisql_ alone.)
Looks good, couple of minor comments on PR (assuming I found a correct PR on github, JIRA is still ignoring qserv PRs).
Andy, can you take a look (next week)?