0

Essentially, I have a dataframe of 300 patients each with 25 recorded variables. I am trying to work out the most common combinations of these variables using the UpsetR package which worked fine for me previously when I did the same for another dataframe with the same number of patients and 14 variables. However, now I keep getting an errors saying (as well as it taking 4hours to compile the code just for me to get this error:


Error: vector memory exhausted (limit reached?)
Error during wrapup: vector memory exhausted (limit reached?)
Error: no more error handlers available (recursive errors?); invoking 'abort' restart

Is this just due to the fact that there is too many combinations/ too many sets that it can't compute the combinations.

Here is the code:


  UpSetR::upset(
    sets = c("a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m", "n", "o","p", "q", "r", "s", "t", "u", "v", "w", "x", "y"),
    order.by = "freq",
    empty.intersections = "on",
    nintersects = 10,
    nsets = 20,
    number.angles = 0,
    point.size = 3.5,
    line.size = 2, 
    mainbar.y.label = "Professionals Combinations",
    sets.x.label = "Patient frequency")

I am a newb with R and havent been able to understand completely what nsets and nintersects and empty.intersections mean from the documentation and unsure if this could have an underpinning to solving my problem. Could it be that I could change some of these functions which might allow me to actually calculate the combinations?

If this doesnt work, how would you try to go around this issue.

Any advice would be great.

bazingastats1203
  • 1
  • 1
  • I have no experience with UpsetR. However, according to this example (https://upset.app/#upset-vs-venn-diagrams) I'd have thought that your 300 patients are the sets (columns) whose elements (rows) are your 25 binary (present/absent) variables. You could thus try first whether the transposed data (25 rows times 300 columns) work. You could also try to feed a polite, say 5 by 10, subset to the online test app at: https://gehlenborglab.shinyapps.io/upsetr/ – I_O Jan 04 '23 at 23:38
  • It may be helpful to post this on their GitHub page. https://github.com/hms-dbmi/UpSetR – Joseph Wood Jan 05 '23 at 23:34
  • You may want to try [`complex-upset`](https://github.com/krassowski/complex-upset) where I did some memory optimizations that might help. – krassowski Jan 07 '23 at 00:45

0 Answers0