scidb (version 1.2-0)

factor_scidb: Hybrid R/SciDB factors

Description

Create R factor variables that include SciDB dimension array index values.

Usage

factor_scidb(x, levels)

Arguments

x

An R vector or factor vector.

levels

A SciDB dimension array. This can be an array created with the SciDB index_lookup function, or any one-dimensional SciDB array. If the array has more than one SciDB attribute, the first attribute will be used to look up values from x.

Value

A factor vector with extra class scidb_factor, and two additional attributes: scidb_levels contains an R vector of looked-up SciDB index values from the levels array, and scidb_index contains a reference to the levels array.

Examples

## Not run: 
# # Consider a SciDB dimension array of Iris flower species, perhaps with
# # some additional data (made up in this example).
# set.seed(1)
# species <- data.frame(
#   species=c("albicans","flavescens","germanica","setosa", "variegata", "versicolor", "virginica"),
#   additional_data = runif(7))
# 
# species <- as.scidb(species, dimlabel="index")
# str(species)
# 
# # Fisher's iris data example contain a subset of these species.
# str(iris$Species)
# 
# # Let's index the iris data with the SciDB array.
# iris$Species <- factor_scidb(iris$Species, species)
# 
# # The Species variable in the data frame now contains *two* indices, the usual
# # R enumeration of levels, and a new set of levels corresponding to the SciDB
# # lookup array. It also contains a reference to the lookup array.
# 
# # Let's upload the newly indexed iris data to SciDB. Observe that the 'Species'
# # values are uploaded as SciDB int64 index values! Those indices are join-able
# # with the dimension array used to create the factor.
# x <- as.scidb(iris)
# str(x)
# 
# # The advantage is that the x$Species values can be joined or redimensioned
# # conformably with the SciDB indexing array.
# 
# # The next example computes the average iris data values grouped by Species
# # using SciDB's redimension function:
# xr <- redimension(x, dim="Species", FUN=mean)
# 
# # That output is join-able with the SciDB dimension array species:
# merge(xr,species, by.x="Species", by.y="index")[]
# 
# # ...the output should look something like this...
# #  Sepal_Length Sepal_Width Petal_Length Petal_Width    species additional_data
# #4        5.006       3.428        1.462       0.246     setosa       0.9082078
# #6        5.936       2.770        4.260       1.326 versicolor       0.8983897
# #7        6.588       2.974        5.552       2.026  virginica       0.9446753
# 
# ## End(Not run)