aggregateDimension: SWS Dimension Aggregation

In SWS-Methodology/faoswsUtil: A suite of function to support the FAO new Statistical Working System

Description

Aggregate the suppled data.table along the named dimension using the supplied aggregation tree.

Usage

aggregateDimension(data, keys, dim_name, key_tree, agg_flags = NULL,
  high_level = 0)

Arguments

data	A long form data.table with the data to be aggregated - result of a faosws::GetData call
keys	A vector containing the data.table key column names (i.e. not the value of flag columns)
dim_name	The name of the dimension to be aggregated
key_tree	The aggregation tree in the form data.table (parent, children)
agg_flags	A data.table containing the flags to attach to aggregated values (optional)
high_level	The highest level to which to aggregate (by default, 0, the highest possible)

Details

Aggregates the data in the supplied data.table according to the supplied key tree. The function works from the bottom of the key tree upwards, cascading the aggregation until it reashes the highest level of the tree structure. The level at which the aggregation stops can be controlled using the high_level parameter. If the agg_flags parameter is supplied, it should contain a data.table with a single row containing the flag(s) to be applied to aggregated values; column names should be the same as those supplied in the data parameter. Any orphans are present as an attribute.

Value

The aggregated rows of data as a data.table

Author(s)

N.A.Connell

Examples

## Not run: 
library(faosws)
library(data.table)
data <- GetData(swsContext.datasets[[1]])
keys <- colnames(data)
keys <- keys[!(keys %in% c("Value", "flagObservationStatus", "flagMethod"))]
key_tree <- GetCodeTree (domain = "agriculture", dataset = "aproduction", 
                         dimension = "measuredItemCPC")
flags <- data.table(flagObservationStatus = c("I"), flagMethod = c("s"))

ag_data <- aggregateDimension(data, keys, "measuredItemCPC", key_tree, agg_flags = flags)

## Another example

flags <- data.table(myFlag = c("A")) 
key_tree <- data.table(
  parent = c(rep("a", 3), rep ("a.1", 3), rep("b", 3)),
  children = c("a.1", "a.2", "a.3",
               "a.1.x", "a.1.y", "a.1.z",
               "b.1", "b.2", "b.3")
)
data <- data.table(
  targetDimension = c("a.2", "a.3",
                      "a.1.x", "a.1.y", "a.1.z",
                      "b.1", "b.2", "b.3", "c.1"),
  otherDimension = c(rep("10", 9)),
  Value = c(1:9),
  myFlag = c(rep("-", 9))
)
keys <- colnames(data)
keys <- keys[!(keys %in% c("Value", "myFlag"))]

## Do the aggregation
aggs <- aggregateDimension(data, keys, "targetDimension",
                           key_tree, agg_flags = flags)

aggs
str(aggs)
attr(aggs, "orphans")

## End(Not run)

SWS-Methodology/faoswsUtil documentation built on May 9, 2019, 11:53 a.m.