send_dataframe should generate a an index column
#8593
Labels
enhancement
New feature or request
feat-dataframe-api
Everything related to the dataframe API
🐍 Python API
Python logging API
Comments
jleibs
added
enhancement
2 tasks
grtlr
added a commit
that referenced
this issue
Apr 2, 2025
### Related * Part of: #8593 * Will unlock: #7204 ### What This brings a very basic table entry to the Rerun viewer. Tables are essentially dataframes, can contain Rerun-native components, but don't have to adhere to some of the restrictions that we place on regular entities (i.e. we don't require indexation columns). <img width="434" alt="image" src="https://github.com/user-attachments/assets/eab906e9-ad62-4052-ac25-dad811f46840" /> ### Testing The following will create a _very_ basic table in-memory. A follow-up PR will add GRPC-based sending of dataframes. ``` RERUN_EXPERIMENTAL_TABLE=1 pixi run rerun ``` ### Known issues * [x] Deleting all recordings switches to welcome screen although tables still present. (tracked via #9465) * [x] Finding the closest entry after closing is not implemented (will be closed via #9464)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
send_dataframeworks nicely if your data is already has a properly annotated index columnHowever, the current implementation implicitly falls back on sending the columns as static data if none of your columns matches an index.
Instead, we should have an option to synthesize a sequential timeline based on row-number so that we get a nice table in the viewer.
The text was updated successfully, but these errors were encountered: