I'm trying to send and execute an R script to the IBM dashDB REST endpoint based on the tutorial.
Everything works quite nice but when I use the following code
timestamp=as.numeric(strptime(df$TIMESTAMP, \"%Y-%m-%d %H:%M:%S.000000\"));
The dashDB REST service simply returns:
An internal error has occurred. Please check the URL and try again.
Using the R web console inside the dashDB service the script runs fine. I've attached the whole script at the end of this message.
Any ideas?
library(ibmdbR)
mycon <- idaConnect("BLUDB", "", "")
idaInit(mycon)
df <- as.data.frame(ida.data.frame('"DASH110683"."TEMPERATURE"')[ ,c('NAME', 'TEMPERATURE', 'TIMESTAMP')])
pattern="%Y-%m-%d %H:%M:%S.000000"
timestamp=as.numeric(strptime(df$TIMESTAMP, pattern))
maxTimestamp=max(timestamp)
cutOffTimestamp=maxTimestamp - 10
idx = which(timestamp>cutOffTimestamp)
workingSubset=df[idx,]
devideIds=unique(workingSubset$NAME)
scores = 1:length(devideIds)
for (i in 1:length(devideIds)) {
idx = which(workingSubset$NAME==devideIds[i])
df=workingSubset[idx,]
values=as.numeric(df$TEMPERATURE)
#scores[i]=(sqrt((globalMean- mean(values))^2) +sqrt((globalSD- sd(values))^2))/2
#scores[i]=abs(globalSD- sd(values))
scores[i]=sd(values)
}
if (max(scores)>1) {
devideIdOutlier=devideIds[scores==max(scores)]
} else {
devideIdOutlier=0;
}
Romeo, the API described in the tutorial is more of a preliminary one. In fact we have just made available the new strategic one that also provides more options. You can now either pass in the R script via a request parameter or you can reference a R script that is stored inside dashDB (e.g. by using the built-in RStudio and saving the script).
https://developer.ibm.com/clouddataservices/wp-content/themes/projectnext-clouddata/dashDB/