table.DownsideRiskRatio/table.Variability variable freq not defined

[ThomasHehehehe]

Hello,
in table.DownsideRiskRatio function:
freq is not defined if scale is not NA.

y = checkData(R)
columns = ncol(y)
columnnames = colnames(y)
if (is.na(scale)) {
freq = periodicity(R)
...
}
for (column in 1:columns) {
...
znames = c(paste0(**freq$**scale, " downside risk"), "Annualised downside risk",
"Downside potential", "Omega", "Sortino ratio",
"Upside potential", "Upside potential ratio", "Omega-sharpe ratio")
...
}

[braverock]

thanks for the report.

Could you describe why you think this is a problem? If you don't pass scale, the function tries to figure it out.

Do you have an example of a problem you're having?

I'm happy to take a look, but I want to know what you were expecting, and what, if anything, is not working.

[ThomasHehehehe]

Thank you for quick response!

table.DownsideRiskRatio(port, scale = 252, digits=4)
does not work
while
table.DownsideRiskRatio(port, digits=4)
works.

I checked the function and found that if scale is defined,
and the code dose not go into if{} ,
therefore freq will not be defined.
However, freq is used anyway in for{} in following line:
znames = c(paste0(freq$scale, " downside risk"), "Annualised downside risk",

Thank you! PerformanceAnalytics is a very nice package.

[ThomasHehehehe]

version: PerformanceAnalytics_1.5.3

[braverock]

upon investigation, your report doesn't make sense to me, and I still dont have a reproducible example.

The code in question is this:

    if(is.na(scale)) {
        freq = periodicity(y)
        switch(freq$scale,
            minute = {stop("Data periodicity too high")},
            hourly = {stop("Data periodicity too high")},
            daily = {scale = 252},
            weekly = {scale = 52},
            monthly = {scale = 12},
            quarterly = {scale = 4},
            yearly = {scale = 1}
        )
    }

so freq is defined if scale is not defined, and it is only used inside that if block, not any where else in the function.

In all cases, scale will be defined, and scale is what is used later in the function.

I suspect that your port object is non-conforming in some way, but without a reproducible example demonstrating the problem you say you're having, I can't demonstrate that.

[ThomasHehehehe]

Hi! Thanks for following up!
Actually, in the script for table.DownsideRiskRatio(),
freq is used outside the if block.

PerformanceAnalytics/R/table.DownsideRiskRatio.R

Lines 61 to 72 in faf539c

	for(column in 1:columns) {
	z = c(DownsideDeviation(y[,column,drop=FALSE], MAR = MAR), DownsideDeviation(y[,column,drop=FALSE], MAR = MAR)*sqrt(scale), DownsidePotential(y[,column,drop=FALSE], MAR = MAR), UpsideRisk(y[,column,drop=FALSE], MAR = MAR, stat = "potential")/DownsidePotential(y[,column,drop=FALSE], MAR = MAR), SortinoRatio(y[,column,drop=FALSE], MAR = MAR), UpsideRisk(y[,column,drop=FALSE], MAR = MAR, stat = "potential"), UpsidePotentialRatio(y[,column,drop=FALSE], MAR = MAR), OmegaSharpeRatio(y[,column,drop=FALSE], MAR = MAR))
	znames = c(paste0(freq$scale," downside risk"), "Annualised downside risk", "Downside potential", "Omega", "Sortino ratio", "Upside potential", "Upside potential ratio", "Omega-sharpe ratio")
	if(column == 1) {
	resultingtable = data.frame(Value = z, row.names = znames)
	}
	else {
	nextcolumn = data.frame(Value = z, row.names = znames)
	resultingtable = cbind(resultingtable, nextcolumn)
	}
	}

In line 64, freq$scale is used.
This causes the bug if scale is well defined in the parameter input.
Hope I make myself clear.
Thank you!

[braverock]

OK, now I understand you. I will fix.

[ThomasHehehehe]

OK, now I understand you. I will fix.

Thank you! Similar issues may also happen in other functions using scale input.

[DzimitryM]

Please fix the same bug in function 'table.Distributions()'

It can be done by just placing the following string outside the block 'if(is.na(scale)) {...}'
freq = periodicity(y)