Send multiple "columns" to HDF5 file?


I am working on an eye tracking experiment and am trying to figure out the best way to send triggers to the recording HDF5 file for later analysis. My goal is to be able to easily find a range of "event id"s in the HDF5 file to extract only tracking data during trials, and also to sort trials depending on the type of stimuli presented.
I have tried ioServer.sendMessageEvent(text=“trial_start”) and similar to send messages, to the file, which works, but as everything ends up in the same column, it is not ideal, but workable I suppose - but I am thinking there is a better way?
I have also tried ioServer.addTrialHandlerRecord(thisTrial) which does give me a copy of the conditions file in the HDF5, but here I don’t get corrosponding "event id"s, which will allow me to easily pick out a range in the eye tracking data.
Does anyone have a good solution?

One approach using the conditions table is to add extra dependent variable columns to hold things like the trial start / end times that are written during experiment runtime. The demos/coder/iohub/eyetracking/gcCursor demo does this. The experiment trial_conditions.xlsx has two columns added for ‘TRIAL_START’ and ‘TRIAL_END’ times, which are written by the experiment at runtime and saved to the hdf5 confitions table for each trial along with any other columns in the conditions file. You can then read the condition variables table, get the trial start and end times written during the experiment, and split events into trials.