Request for flag_paramval

[maowerner]

I would like to have a function in analogy to paramval_filter. The difference would be that instead of removing the rows where the filter condition is TRUE, it would introduce a new parameter that is TRUE or FALSE allowing to filter afterwards.

I need that for the phase shift values. We filter some points out of the Breit-Wigner fits, but plot the unfiltered data to show them as well. I want a column use_for_shift_fit which I could visualize with alpha. My current workaround is

.func <- function(param, value){
  if(
    (param$ensemble == 'A40.20' & param$total_momentum_d_sq == 3 & param$irrep == 'A1' & param$correlator_id == 2) |
    (param$ensemble == 'A40.32' & param$total_momentum_d_sq == 2 & param$irrep == 'B1' & param$correlator_id == 2) |
    (param$ensemble == 'A40.32' & param$total_momentum_d_sq == 4 & param$irrep == 'E'  & param$correlator_id == 2) |
    (param$ensemble == 'B35.32' & param$total_momentum_d_sq == 3 & param$irrep == 'A1' & param$correlator_id == 3) |
    (param$ensemble == 'B55.32' & param$total_momentum_d_sq == 3 & param$irrep == 'E'  & param$correlator_id == 1) |
    (param$ensemble == 'D15.48' & param$total_momentum_d_sq == 4 & param$irrep == 'E'  & param$correlator_id == 2) |
    (param$ensemble == 'D30.48' & param$total_momentum_d_sq == 1 & param$irrep == 'E'  & param$correlator_id == 2)
    ){
      use_for_shift_fit <- FALSE
    }
  else{use_for_shift_fit <- TRUE}

  value$use_for_shift_fit = use_for_shift_fit

  value
}

followed by

.func <- function(param, value){
  if(value$use_for_shift_fit){ NA }

  value
}

[maowerner]

The second alternative would be to create a new dataframe with this additional parameter. In this case it is a hassle though because I would have to write out ~300 parameter combinations to filter out 10.

[martin-ueding]

Is there still a need to do something here or did you get it resolved some other way?