idxv: Precalculate data.table indices

In jangorecki/dwtools: Data Warehouse related functions.

Description

Precalculate indices for data.table, also known as Nth key.

Usage

idxv(DT, Idx, grp = FALSE)

Arguments

DT	data.table for which generate indices.
Idx	list of character vectors, indices to generate.
grp	logical (default FALSE), when TRUE it will include groups id which can be used to speedup aggregations.

Value

list of data.table, including some meta data in attributes.

See Also

CJI

Examples

suppressPackageStartupMessages(library(dwtools))

DT = dw.populate(scenario="fact")
setkeyv(DT,c("cust_code", "prod_code", "geog_code", "time_code", "curr_code"))
print(DT)

Idx = list(
  c("cust_code", "prod_code", "geog_code"),
  c("cust_code", "geog_code", "curr_code"),
  c(2:3)
)
IDX = idxv(DT, Idx)
# list indices
attr(IDX,"DT.key")
# all indices size in MB
cat("Total indices count: ",length(IDX),", size: ",tables(env=as.environment(setNames(IDX,1:length(IDX))),silent=TRUE)[,sum(as.numeric(MB))]," MB\n",sep="")

# Idx[[1]] # filter
DT[CJI(IDX,"id020",847,"AK")]
DT[cust_code=="id020" & prod_code==847 & geog_code=="AK"]
# Idx[[2]] # filter
DT[CJI(IDX,"id006",TRUE,"UT",TRUE,"NOK")]
DT[cust_code=="id006" & geog_code=="UT" & curr_code=="NOK"]
# Idx[[3]] # filter
DT[CJI(IDX,TRUE,323,"OR")]
DT[prod_code==323 & geog_code=="OR"]

# verify DT key
key(DT) # still the same

jangorecki/dwtools documentation built on May 18, 2019, 12:24 p.m.