fillmap: Supplement missing values in mapping of data

In preputils: Utilities for Preparation of Data Analysis

Description

In properly normalized data bases, 1:1 mapping should be complete and unique. In real world data however ID mappings or data base key candidates are repeated over and over across observations, partly with missing data in case of merged data set. fillmap supplements NAs in mapping variables as far as possible

Usage

fillmap(x, y, what = "xy", rmdup=FALSE, rmmiss=FALSE,
    printori=FALSE)

Arguments

x,y	vectors of equal length, holding the mapping values, sepearated by ","
what	data to be returned, either 1st ("x") or 2nd argument ("y") or a data,table, cointaining both ("xy")
rmdup	remove duplicates from mapping (TRUE) or return all rows in original order (FALSE)
rmmiss	remove rows, where not mapping could be found (TRUE) or return all rows (FALSE)
printori	print original variables side by side

Details

incons assumes a 1:1 mapping between provided variables, as is commonly the case for example in ID translation steps.

For all cases where a proper unambiguous 1:1 matching exists. The missings values are filled in

Value

Vector or data.table with original mapping data, where NAs are filled in whith supplemented data where possible

Examples

    library(data.table)
    pheno1 <- data.frame(id1=c(1,2,3,4),id2=c(11,22,NA,NA),phenodat=c(NA,NA,NA,"d"))
    pheno2 <- data.frame(id1=c(NA,NA,NA),id2=c(11,22,33),phenodat=c("a","b","c"))
    pheno3 <- data.frame(id1=c(4,3),id2=c(44,33),phenodat=c(NA,NA))
    phenoges <- rbind(rbind(pheno1,pheno2),pheno3)
    with(phenoges,fillmap(id1,phenodat))
    with(phenoges,fillmap(id1,phenodat,rmdup=TRUE))
    with(phenoges,fillmap(id1,phenodat,rmmiss=TRUE))
    with(phenoges,fillmap(id1,phenodat,rmdup=TRUE,rmmiss=TRUE))
    with(phenoges,fillmap(id2,phenodat))
    with(phenoges,fillmap(id2,phenodat,rmdup=TRUE))
    with(phenoges,fillmap(id2,phenodat,rmmiss=TRUE))
    with(phenoges,fillmap(id2,phenodat,rmdup=TRUE,rmmiss=TRUE))
    phenosupp <- with(phenoges,fillmap(id1,id2))
    names(phenosupp) <- c("id1","id2")
    phenosupp$phenodat <- fillmap(phenosupp$id1,phenoges$phenodat,what="y")
    unique(phenosupp)

preputils documentation built on July 1, 2020, 5:35 p.m.