Flow Support Forum

Hi, I am trying out the csv viewer and got quite quick results, loving the way it autodetects groups etc.

However I am a bit confused about the handling of x-axis. Even though I have enabled saving the setting for variables in the project files, the setting of using a variable as an x-axis is not saved, I always have to after opening reset the x axis for the variable I want. Also I am using seconds since epoch as a float for it, is there a way that it is automatically converted into a proper (utc?) timestamp.

PS: it would also be very helpful if one could autocomplete variable names in the formula editor, or at least copy them from elsewhere, ours get really complicated.

Hello,

You're correct, Flow doesn't store the X axis settings, which clearly it should. I'll fix it for the next version. Sorry about the inconvenience!

Flow doesn't have a datetime type, but as a matter of fact there is an undocumented function for turning a Unix timestamp (milliseconds since Jan 1, 1970) into a string of the format yyyy-mm-dd hh:mm:ss. The name of the function is str_datetime. So if your time variable is time (in seconds since epoch), you'd write: Note that the name of this function may change in future version - that's why it's undocumented.

In Flow CSV Editor, the paid version, it should be possible to double-click a variable name and copy it, so you can paste it into your formula afterwards. In Flow CSV Viewer, this isn't possible, but a workaround could be to open the CSV file and copy the names from there. Not ideal obviously, but possibly helpful... Anyway, autocompletion of variable names sounds like a great idea, so I'll add it to the list. Thanks!

Best regards,
Gustaf

Thank you very much, that datetime is very helpful and I guess a lot of people would not mind it being an official feature.

Another question regarding x axis: Is there a way to make something a "true x axis" and not just display the values? I have lots of data where the time between points varies a lot (e.g. for saving space two equal points would not be output) and it makes interpretation of the data somewhat difficult.

Glad it helped! No unfortunately not, and that's by far the most requested feature - quite understandably. It's a bit complicated to implement however, because the entire caching mechanism that makes Flow quite fast to plot big data files depends on a fixed data point width. But it will come, eventually! I'll not yet another request for this.

Best regards,
Gustaf