Opasnet Base Connection for R

This page is a knowledge crystal of subtype method. The page identifier is Op_en4614
Moderator:Teemu R (see all)
Upload data {{#opasnet_base_link:Op_en4614}}

Code for R for the purpose of interacting with the Opasnet Base is collected on this page. To use it, copy paste the code to R; this defines the functions, after which they can be called in R. I've tried to explain the code in comments (#). This is a work in progress, if you have comments or new ideas post them using the comment templates on a new line after the one you want to comment, I'm no expert in programming so input is appreciated :P.

Package dependencies

These packages are required for most of the code to work. Load these first.

1. utils
2. RODBC

Setup

Establishes the connection to the database. Copy paste first.

db <- odbcConnect("opasnet_base") #assuming "opasnet_base" is the defined DSN

Downloading data

• Tested and mostly works, in case of error with GetArray try GetDataFrame, it should work with data missing indexes and locations.

Low level functions (mainly sql queries)

Copy paste these second.

#Get latest series id
op_baseGetLatest <- function(ident) {
	series_list <- sqlQuery(db, paste('
		SELECT actobj.series_id 
		FROM obj 
		LEFT JOIN actobj 
		ON obj.id = actobj.obj_id 
		WHERE obj.ident = "', ident, '"', sep = ''))
	series_list[nrow(series_list),1]
}

#Get results (cell id, result, obs)
op_baseGetData <- function(page_ident, series_id) {
	sqlQuery(db, paste('
		SELECT cell.id, res.result, res.obs 
		FROM obj 
		LEFT JOIN actobj 
		ON obj.id = actobj.obj_id
		LEFT JOIN cell 
		ON actobj.id = cell.actobj_id 
		LEFT JOIN res 
		ON cell.id = res.cell_id 
		WHERE actobj.series_id = ', series_id, 
		' AND obj.ident = "', page_ident, '"', sep = ''))
}

#Not in use: Get indexes and locations per cell
op_baseGetDims <- function(cell_id_list) {
	sqlQuery(db, paste('
		SELECT loccell.cell_id, loc.id, loc.location, loc.obj_id_i, obj.name 
		FROM loccell 
		LEFT JOIN loc 
		ON loccell.loc_id = loc.id 
		LEFT JOIN obj 
		ON loc.obj_id_i = obj.id 
		WHERE loccell.cell_id BETWEEN ', cell_id_list[1,1], ' AND ', cell_id_list[nrow(cell_id_list),1], sep = ''))
}

#Experimental: match locations and indexes
#tapply(dims_table[,3], dims_table[,5], paste)

#Not in use: Get indexes from dims_table
op_baseFindInd <- function(dims_table) {
	factor(op_baseDims[,4]) 
}

#Get indexes
op_baseGetInd <- function (cell_id_list) {
	indexes <- sqlQuery(db, paste('
		SELECT DISTINCT loc.obj_id_i
		FROM cell 
		LEFT JOIN loccell 
		ON cell.id = loccell.cell_id
		LEFT JOIN loc 
		ON loccell.loc_id = loc.id 
		WHERE cell.n <> 0 AND
		cell.id BETWEEN ', cell_id_list[1,1], ' AND ', cell_id_list[nrow(cell_id_list),1],
		' ORDER BY loc.obj_id_i', sep = ''))
	as.numeric(indexes[,1])
}

#Get locations
op_baseGetLoc <- function(index, cell_id_list) {
	locations <- sqlQuery(db, paste('
		SELECT DISTINCT loc.id 
		FROM loccell 
		LEFT JOIN loc 
		ON loccell.loc_id = loc.id 
		WHERE loc.obj_id_i = ', index, ' AND loccell.cell_id BETWEEN ', cell_id_list[1,1], ' AND ', 
		cell_id_list[nrow(cell_id_list),1], sep = ''))
	as.numeric(locations[,1])
}

#Not in use: Get location names
op_baseGetLocNames <- function(index, cell_id_list) {
	locations <- sqlQuery(db, paste('
		SELECT DISTINCT loc.location, loc.obj_id_i 
		FROM loccell 
		LEFT JOIN loc 
		ON loccell.loc_id = loc.id 
		WHERE loc.obj_id_i = ', index, ' AND loccell.cell_id BETWEEN ', cell_id_list[1,1], ' AND ', 
		cell_id_list[nrow(cell_id_list),1], sep = ''))
	as.numeric(locations[,1])
}

#Get location name (simple)
op_baseGetLocName <- function(loc_id) {
	sqlQuery(db, paste('
		SELECT location 
		FROM loc 
		WHERE id = ', loc_id, sep = ''))
}

#Get index name
op_baseGetIndName <- function(ind_id) {
	sqlQuery(db, paste('
		SELECT name 
		FROM obj 
		WHERE id = ', ind_id, sep = '')) 
}

#Get locations for a specific cell
op_baseGetLocs <- function(cell_id) {
	sqlQuery(db, paste('
		SELECT DISTINCT loc.location, loc.id, loc.obj_id_i 
		FROM loccell 
		LEFT JOIN loc 
		ON loccell.loc_id = loc.id 
		WHERE loccell.cell_id = ', cell_id,
		' ORDER BY loc.obj_id_i', sep = ''))
}

#Get cell.n (cell id, cell.n)
op_baseGetCelln <- function(cell_id_list) {
	sqlQuery(db, paste('
		SELECT cell.id, cell.n
		FROM cell 
		WHERE cell.id 
		BETWEEN ', cell_id_list[1,1], ' AND ', cell_id_list[nrow(cell_id_list),1], sep = ''))
}

#Get results from probabilistic data (cell id, result, obs)
op_baseGetDataProb <- function(page_ident, series_id) {
	sqlQuery(db, paste('
		SELECT cell.id, res.result, res.obs 
		FROM obj 
		LEFT JOIN actobj 
		ON obj.id = actobj.obj_id
		LEFT JOIN cell 
		ON actobj.id = cell.actobj_id 
		LEFT JOIN res 
		ON cell.id = res.cell_id 
		WHERE actobj.series_id = ', series_id, 
		' AND obj.ident = "', page_ident, '"', 
		' AND res.obs <> 0', sep = ''))
}

High Level function

Copy paste this last.

op_baseGetArray <- function(page_ident) {
	
	#Basic data downloading
	
	op_baseLatest <- op_baseGetLatest(page_ident)
	op_baseData <- op_baseGetData(page_ident, op_baseLatest) #cell,result, obs
	
	#Differentiate between probabilistic and non-probabilistic
	
	probabilistic <- FALSE #variable indicator of whether the object is probabilistic
	for (i in 1:length(op_baseData)) {
		if (op_baseData[i,3] > 0) probabilistic <- TRUE
	}
	if (probabilistic == TRUE) op_baseData <- op_baseGetDataProb(page_ident, op_baseLatest) #cell,result, obs
	
	#final_array <- op_baseData[,2] #ideal case (data in proper order, achievable by tweaking upload order)
	#op_baseDims <- op_baseGetDims(op_baseData) #with further understanding of R number of queries could be reduced
	#op_baseIndexes <- op_baseFindInd(op_baseDims) #improved code here
	op_baseInd <- op_baseGetInd(op_baseData)
	if (probabilistic == TRUE) op_baseCelln <- op_baseGetCelln(op_baseData)
	
	#Array structure building
	
	IndLengths <- 1:3 #blank vector for index lengths
	IndList <- rep(vector("list", 1), length(op_baseInd)) #blank list of appropriate length for indexes and 
	#locations for naming the dimensions of the array
	IndIdList <- rep(vector("list", 1), length(op_baseInd)) #blank list of appropriate length for ids of indexes
	IndNames <- c(1) #blank vector for index names
	for (i in 1:length(op_baseInd)) {
		IndIdList[[i]] <- op_baseGetLoc(op_baseInd[i], op_baseData) #fetch location ids in index to a list, 
 		#assumed every cell has locations for all unique index id values in loc.obj_id_i in the range of cells
		IndNames[i] <- paste(op_baseGetIndName(op_baseInd[i])[1,1]) #fetch index name to a vector
		IndLengths[i] <- length(IndIdList[[i]]) #backup: op_baseGetLoc(op_baseInd[i], op_baseData))
		for (j in 1:IndLengths[i]) { 
			IndList[[i]][j] <- paste(op_baseGetLocName(IndIdList[[i]][j])[1,1])
		}
	}
	if (probabilistic == TRUE) { #Probabilistic exceptions (add index: Iteration)
		Iterations <- max(op_baseCelln[,2])
		IndLengths[length(IndLengths) + 1] <- Iterations
		IndList[[length(IndList) + 1]] <- 1:Iterations
		IndNames[length(IndNames) + 1] <- "Iteration"
	}
	final_array <- rep(NA, prod(IndLengths))
	dim(final_array) <- IndLengths
	names(IndList) <- IndNames
	names(IndLengths) <- op_baseInd
	dimnames(final_array) <- IndList
	
	#Cell population
	
	IndexNumber <- 1:length(op_baseInd)
	names(IndexNumber) <- op_baseInd
	LocList <- c(1) #blank vector for locations per cell
	LocMap <- c(1) #blank vector for location to position mapping
	LocMapNames <- c(1)
	CellPosition <- rep(1, length(IndLengths)) #blank vector for cell position
	dim(CellPosition) <- c(1, length(IndLengths)) #form into matrix of appropriate form
	for (i in 1:length(op_baseData[,2])) { #for each result
		LocList <- op_baseGetLocs(op_baseData[i,1]) #get list of locations
		for (j in 1:length(LocList[,1])) { #for each location per cell
			LocMap <- 1:IndLengths[paste(LocList[j,3])] #get available positions in given index 
			names(LocMap) <- dimnames(final_array)[[paste(names(IndList)[IndexNumber[paste(LocList[j,3])]])]] 
			#add names (locations) of positions in given index
			CellPosition[1,IndexNumber[paste(LocList[j,3])]] <- LocMap[paste(LocList[j,1])] 
			#match cell specific locations to positions in index
		}
		if (probabilistic == TRUE) CellPosition[1,length(CellPosition)] <- op_baseData[i,3]
		final_array[CellPosition] <- op_baseData[i,2] #insert result in array in position determined by CellPosition
	}
	final_array #print/return the final array
}

ArraytoDataFrame <- function(array) {
	as.data.frame(as.table(array))
}

op_baseGetDataFrame <- function(page_ident) {
	
	#Basic data downloading
	
	op_baseLatest <- op_baseGetLatest(page_ident)
	op_baseData <- op_baseGetData(page_ident, op_baseLatest) #cell,result, obs
	
	#Differentiate between probabilistic and non-probabilistic
	
	probabilistic <- FALSE #variable indicator of whether the object is probabilistic
	for (i in 1:length(op_baseData)) {
		if (op_baseData[i,3] > 0) probabilistic <- TRUE
	}
	if (probabilistic == TRUE) op_baseData <- op_baseGetDataProb(page_ident, op_baseLatest) #cell,result, obs
	
	op_baseInd <- op_baseGetInd(op_baseData)
	
	#Table building
	
	for (i in 1:length(op_baseData[,2])) { #for each result
		LocList <- paste(op_baseGetLocs(op_baseData[i,1])[,1]) #get list of locations
		if (length(LocList) != 0) {
			for (j in 1:length(LocList)) { #for each location per cell
				op_baseData[i,(3+j)] <- LocList[j] #put in place in table
			}
		}
	}
	final_table <- data.frame(1:length(op_baseData[,1]))
	#final_table[,1] <- op_baseData[,1]
	x <- 0
	if (probabilistic == TRUE) {
		x <- 1
		dimnames(final_table)[[2]][1] <- "Iteration"
		final_table[,1] <- op_baseData[,3]
	}
	if (is.na(op_baseInd[1]) == TRUE) y <- 0 else y <- length(op_baseInd)
	if (y > 0) {
		for (i in 1:y) {
			final_table[,x + i] <- op_baseData[,3+i]
		}
	}
	final_table[,(x + y + 1)] <- op_baseData[,2]
	if (y > 0) {
		for (i in 1:y) {
			dimnames(final_table)[[2]][x+i] <- paste(op_baseGetIndName(op_baseInd[i])[1,1])
		}
	}
	dimnames(final_table)[[2]][(x + y + 1)] <- "Result"
	final_table
}

Usage

variable <- op_baseArray("<page identifier>")
variable <- ArraytoDataFrame(<array name>)
variable <- op_baseGetDataFrame("<page identifier>")