SD score used instead of measurements when evaluating "extraneous-same-day" for the last time (step 12d)

[delosh653]

In lines 638 - 662 of growth.R (step 12d), it uses the SD score when it should be the actual measurements:

  #  ii.  Determine dup_tot_* (# of days with extraneous for that subject/parameter) and nodup_tot_* (# of days with nonexlcuded
  #       non-extraneous for that subject/parameter).
  # iii.  If dup_tot_*/(dup_tot_*+nodup_tot_*) is greater than 1/2, replace exc_*=7 for all extraneous for that subject/parameter
  #       for each age where the  largest measurement minus the smallest measurement for that subject/parameter/age is larger than
  #       the maximum difference (ht 3cm; wt 0-9.999 kg 0.25kg; wt 10-29.9999 kg 0.5 kg; wt 30kg and higher 1 kg).

  data.df[J(dup.ratio.df[dup.ratio > 1 / 2, list(subjid, param)]), exclude := (function(df) {
    df[, `:=`(tbc.sd.min = as.double(NaN),
              tbc.sd.max = as.double(NaN))]
    df[valid(
      exclude,
      include.extraneous = T,
      include.temporary.extraneous = T
    ), `:=`(tbc.sd.min = min(tbc.sd),
            tbc.sd.max = max(tbc.sd))]
    df[tbc.sd.max - tbc.sd.min > ifelse(param == 'HEIGHTCM',
                                        3,
                                        ifelse(
                                          param == 'WEIGHTKG',
                                          ifelse(tbc.sd.min < 10, 0.25, ifelse(tbc.sd.min < 30, 0.5, 1)),
                                          NA
                                        )),
       exclude := 'Exclude-Extraneous-Same-Day']
    return(df$exclude)
  })(copy(.SD)), .SDcols = c('exclude', 'tbc.sd'), by = .(subjid, param, agedays)]

It does something similar in step 12e. We may want to reevaluate this step overall anyway.

[dchud]

Note the reference to ticket 33 above is incorrect.