split results to improve responsiveness

[edward-burn]

I have a result object of 900,000+ rows. At the moment when we create tables in the shiny app we point to the full data, but then for every time we create a table there is time spent filtering to the result type of interest. I think if in global the data was split into a list with one object per result type then the responsiveness of the shiny app should be much improved. Below for example if I have already filetered the result type, the observation table takes 0.5 seconds to create rather than over 2 seconds

tictoc::tic()
data |> 
  tableObservationPeriod()
tictoc::toc() # 2.07 sec elapsed


data_obs_period <- visOmopResults::filterSettings(data, result_type == 
                                           "summarise_observation_period")

tictoc::tic()
data_obs_period |> 
  tableObservationPeriod()
tictoc::toc() # 0.5 sec elapsed