graphs auto transpose

[Neil.Pate]

When doing quick n dirty prototyping I end up using 2D arrays a lots of the time, often with quite a bit of data in them (very often more than 10k samples per channel).

Dropping a graph onto the FP and then wiring in the data and running the VI, I very often get LV grinding to a halt as it is expecting my channels as rows and I have them as columns etc. From that point on every mouse click is a painfull process to sit though as I have to tell the graphs to transpose the data.

What I propose is that LV looks at the size of the input data, and if one of the indices is much greater than the other then it figures out whether its rows or columns as the data channels. Who ever plots 10k channels each containing only 4 samples?

[David_L]

I've seen a lot of different crazy things done in LabVIEW and one thing I can assure you is this... There absolutely IS someone out there who wants to plot 10K channels with 4 samples each.

[JackDunaway]

I have used graphs for non-standard purposes, such as plotting many traces with only 2 points each. That's not as ridiculous as 10k plots @ 4 samples, but it still bucks the typical ratio where num plots < num points. If this feature is implemented, I would want it off by default (which, by nature, decreases the efficacy and spirit of this Idea...) so that it doesn't break current code.

If you run into a scenario where it takes a long time to regenerate the data (i.e., it's not as straightforward as another Run arrow and 1sec of execution), go ahead and copy the graph indicator onto a new VI (which brings the data along with it also). Turn that new indicator into a control, perform the appropriate transformations, and write it back out to a new indicator. Then, with the new indicator, "Copy Data" and then "Paste Data" on the original VI's indicator - we're talking 20 seconds here. I'll commonly do that with arrays that I biff up and want to transpose or post-process. You can also right-click on the new control and select "Make Current Value Default" and save the VI for later processing - a quick and dirty method of not needing to write test results to a file or DB. It's not quite as quick as the proposed Idea, but I'd still classify it as quick and dirty.

(I'm sure most people reading this are saying "no duh" to the previous paragraph - it's just for the benefit of those not as versed in the art of quick and dirty. )

Message Edited by JackDunaway on 04-13-2010 06:40 PM

[Intaris]

IF this is implemented, it should only occur upon wiring up the graph terminal, not on receipt of data, that would be terrible.

[Neil.Pate]

This is not something I picture happening often, i.e. you would not usually run a VI with this enabled, it is purely a development tool.All I want is some automatic way to stop the PC locking up (99% work unsaved for 10 mins+) when a silly number of channels is attempted to be plotted.

 

When I said 10k I mean more like 2M, that really, really causes the PC to get upset! David_L, know anyone who plots 2M channels of 4 samples? 😉

[Darren]

Any idea that has received less than 3 kudos within 3 years after posting will be automatically declined.