slice and slab ignore overlapping horizons

[dylanbeaudette]

What is the correct output from slice and slab in the presence of overlapping horizons?

Overlapping horizons:

• data-entry error, current strict behavior is ideal
• "A/E" type horizon or multiple lab samples per genetic horizon: possibly wrong output

1: In get.slice(h, id = id, top = top, bottom = bottom, vars = vars,  :
  Bad horizonation detected, first matching horizon selected. Use strict=TRUE to enforce QA/QC.

Relevant notes from get.slice().

• this function cannot deal with overlapping horizons (usually an error): > 1 row / slice
• it would be useful to support these kind of data, as many lab-sampled sites have sub-samples

Example.

library(aqp)

# overlapping horizons by design
d <- data.frame(
  id='001',
  name=c('A', 'A/E', 'A/E', 'Bhs1', 'Bhs2', 'C1', 'C2'),
  top=c(0,5,5,10,20,50,80),
  bottom=c(5,10,10,20,50,80,100),
  var=c(15, 4, 8, 15, 18, 2, 3),
  v2=1,
  stringsAsFactors=FALSE
  )

# init SPC
depths(d) <- id ~ top + bottom

# error due to overlapping horizons
slice(d, 0:50 ~ ., strict=TRUE)

# warnings
s <- slice(d, 0:50 ~ ., strict=FALSE)

# should there be 2 rows of data here ?
# only the first is returned
s[1, 8]

[dylanbeaudette]

Closing this, can re-open as needed for re-written slab and dice