For the Lasair broker, this represents a qualitative difference: (1) is simply a browsable list of attributes, but for (2), we would split it into a “core” list of more useful attributes plus ~150 “extended” attributes that users won’t see at first sight – the design of the web interface would change significantly.
Is the intention to grow the diaObject schema to triple its current number? In other words, which model should we be working toward, (1) or (2)?
Hi @roy, quick note from me as Forum moderator to say that since this looks like a question for Rubin staff about planned data products (in this case, diaObject table schema), I’ve moved it to the “Support - Data Products” category so Rubin staff can follow-up.
(The original category, “Science - Alerts & Brokers” is also relevant to the topic but is for community discussion and is not monitored by Rubin staff.)
Hi @roy, appreciate the pointer. The alert schemas are in general supposed to track the APDB schemas more or less exactly. We had excluded some of the DIAObject timeseries features from the alerts while we worked on determining the feature set, but this was and is intended to be temporary. We will update the schemas accordingly, although I’m not certain if we will do so before OR5. I would indeed expect the number of columns in the DIAObject table to be closer in number to that of the APDB.