Fork me on GitHub

Are vectors supposed to be acceptable as xt/id's? "The only constraint XTDB places on entity ids (:xt/id) is that they must be unique. They can be of any type" suggests vectors are OK (that's on, but vectors are not on the list of "valid IDs" in


Ah, that first answer is a mistake sorry, the second link is definitely correct 🙂 thanks for mentioning this


xtdb core2 is to supposed to supersede current xtdb right ?


Hey @U45SLGVHV. core2 is a research project, whereby ideas may be brought to XTDB, and to explore new areas entirely. Core2 is not currently planned to be a succeeding version of XTDB.


I thought it was meant to replace xtdb as well. Maybe a bit misleading name?


Yes I think it possibly is! We will be aiming to straighten this out soon, to avoid confusion.


personally since core2 is exploring transition from datalog to sql first I’ve avoided using xtdb 1 as I don’t want to invest in project that is not aligned with where vision is headed


It's good that i asked the question then 🙂 Looks like a lot of people are thrown off by the naming and versioning of the project. @U050DD55V thanks for letting me know, Thanks for XTDB btw it's awesome


being “thrown off” implies that xtdb direction is not changing, but likelihood is it will be if core2 experiments go well. so if anything the impressions are accurate, just not confirmed. and it’s unclear what level of support/improvement the datalog version will get if so.


so there is some Osborne effect here, that people don’t want to start using XTDB until there is clarity, I guess


I, for one, don’t want to stop using the current Datalog XTDB and switch to a SQL based


We do plan to deosbornify in the new year, please bear with us. Core2 is currently a research project, with dynamic SQL being just one element - others being separation of storage from compute, columnar storage, and new temporal functionality. We will be focusing now on what this means for XTDB proper, and we will give updates. In the meantime development on XTDB has continued with major performance work, and this is the immediate short term focus along with deeper work on checkpointing, as we look to circle back the R&D from Core2.

👍 1