Log and Ingest
How-to guides for sending data to Rerun.
-
Send entire columns at once
Log many rows of data in a single call
-
Send dataframes to a recording
Log a whole table into a recording in a single call
-
Using layers to append data to segments
Append data to existing segments via named layers
-
Send partial updates over time
Update only the components that change between frames
-
Send tables to Rerun
Send tables as dataframes to the Rerun Viewer
-
Share recordings across multiple processes
Log to a single recording from multiple processes, threads, machines, or languages
-
Log to several recordings from one process
Create more than one recording from a single process, and why calling rr.init again replaces the global recording
-
Send user-defined data
Log custom data via AnyValues and DynamicArchetype
-
Log Eigen, OpenCV, and custom C++ types without copying
Log data from Eigen, OpenCV, or your own C++ containers, with or without a copy, using rerun::Collection and CollectionAdapter
-
Clear out data using tombstones
Mark data invalid after a point in time
-
Working with MCAP
Open and inspect MCAP files in Rerun
-
Working with LeRobot datasets
Open and convert LeRobot datasets in Rerun
-
Optimize chunk count
Reduce chunk count to improve ingest and query performance
-
Work with .rrd files
Inspect, migrate, trim, split, merge, compact, and append to .rrd files with the rerun rrd command
-
Loading URDF models
Load and visualize robot models defined in URDF