Hi, I have a Pavlovia experiment that I created in a previous version of the Builder in 2020 and it worked fine then but now is creating comically bloated data files after updating it to the current version of Builder/Psychopy/Pavlovia. For each trial in the main loop (~500 trials, but just one loop), the script takes a line from a csv, displays an image that was specified on that line (e.g. “square.jpg”, “circle.jpg”) and collects a response. In the 2020 version it created nice, compact data files, but now it is creating data files with five unique columns for each stimulus, that are completely empty for most trials (e.g., for the square.jpg item, it creates square.thisRepN, square.thisTrialN, square.thisN, square.thisIndex, square.ran; for circle.jpg it creates circle.thisRepN, circle.thisTrialN, circle.thisN, circle.thisIndex, circle.ran), and it does that for each of the 500 items. I do not understand why the script would now be creating unique columns for each stimulus, nor how to prevent it. Is there a setting in the new version that I need to tick to tell psychopy/pavlovia to re-use the same five columns for each stimulus in the loop, instead of creating new columns each time? Thanks!
If you have a column called circle.thisN please could you double check for a loop or trial handler called circle?
Thanks for your response. No, there’s no loop or trial handler called ‘circle’. What I finally discovered was that the conditions csv contained a column called ‘name’ with a name for the stimulus. With the recent updates to Psychopy/Pavlovia and how it handles csv conditions files, ‘name’ has apparently become a internally designated variable name in some way, so importing a conditions csv with a column called ‘name’ causes P/P to override its own name for the loop or trial handler when writing the data file.
1 Like