denormalize: Create a (redundant) horizon-level attribute from a...

View source: R/denormalize.R

denormalizeR Documentation

Create a (redundant) horizon-level attribute from a site-level attribute


Create a (redundant) horizon-level attribute from a site-level attribute. Specify a SoilProfileCollection and a site-level attribute from that SPC (by name) to receive a vector of length equal to the number of horizons containing the site-level values. This vector is directly usable with the SoilProfileCollection horizon setter.

denormalize is the inverse operation for the formula interface that "normalizes" a horizon level variable to site level:

site(object) <- ~ horizonvar


denormalize(object, attr)



A SoilProfileCollection


Site-level attribute name (character string) to denormalize to horizon.


"Denormalization" is the process of trying to improve the read performance of a database, at the expense of losing some write performance, by adding redundant copies of data or by grouping data. Sometimes it is beneficial to have site-level attributes denormalized for grouping of horizon-level data in analyses. denormalize achieves this result for SoilProfileCollections.


A vector of values of equal length to the number of rows in the horizon table of the input SPC.


Andrew G. Brown, Dylan Beaudette



# create a SoilProfileCollection from horizon data
depths(sp1) <- id ~ top + bottom

# create random site-level attribute `sitevar` with a binary (0/1) outcome
sp1$sitevar <- round(runif(length(sp1)))

# use denormalize() to create a mirror of sitevar in the horizon table
# name the attribute something different (e.g. `hz.sitevar`) to 
# prevent collision with the site attribute
# the attributes can have the same name but you will then need 
# site() or horizons() to access explicitly
sp1$hz.sitevar <- denormalize(sp1, 'sitevar')

# compare number of profiles to number of sitevar assignments

# compare number of horizons to number of horizon-level copies of sitevar `hz.'sitevar`

aqp documentation built on Sept. 8, 2023, 5:45 p.m.