IMPORTANT: Version 1.80 develop branch now on new python/pandas versions and does not depend on arctic #1293
robcarver17
announced in
Announcements
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Python is 3.10, and pandas is 2.1.3. Default storage for time series is parquet, as per the recent announcement you will need to copy your data to parquet and install pyarrow.
The code for arctic is still present, but we don't use it, so if you haven't copied your data across you will be faced with things breaking / empty data sets if you run this version.
I've run this for a few days succesfully in production, but there may still be crevices of eg sysinit code that rely on out of date pandas API calls.
The only slight gotcha is things break on the most up to date pymongo (4.*), but it's very easy to rewind to 3.11.3 which is what is required.
As noted in the other discussion, long term plans would be to create the possibility of having a pure parquet build which didn't require any server to be run, and to ressurect support for arctic once it's clearer what the situation is there. Basically I'd like to end up in a situation where there are a number of choices of data storage.
Beta Was this translation helpful? Give feedback.
All reactions