Error when reading optimize_method in optimize.portfolio.rebalancing

[goodnewz]

The new CVXR method introduced a bug when reading the optimize_method user input. This happens if the uses CVXR and defines an optimizer (e.g., c("CVXR", "SCS")). The expectation is that optimized_method is a character vector of length 1. However, if the user includes an optimizer, it becomes a character vector of length 2.

Example code:

library(PortfolioAnalytics)
library(CVXR)
library(data.table)
library(xts)

data(edhec)
# Use edhec for a returns object
ret_edhec <- tail(edhec, 60)
colnames(ret_edhec) <- c("CA", "CTAG", "DS", "EM", "EMN", "ED", "FIA",
                         "GM", "LSE", "MA", "RV", "SS", "FF")
print(head(ret_edhec, 5))


fund_edhec <- colnames(ret_edhec)

# Create portfolio object
pspec_maxret <- portfolio.spec(assets=fund_edhec)
# Add constraints to the portfolio object
pspec_maxret <- add.constraint(pspec_maxret, type="full_investment")
pspec_maxret <- add.constraint(portfolio = pspec_maxret, type = "box",
                               min = rep(0.02, 13),
                               max = c(rep(0.6, 4), rep(0.2, 9)))
# Add objective to the portfolio object
pspec_maxret <- add.objective(portfolio = pspec_maxret,
                              type = "risk", name = "StdDev")
pspec_maxret <- add.objective(portfolio = pspec_maxret,
                              type = "return", name = "mean")
pspec_maxret

# Run the optimization with default solver
opt_maxret <- optimize.portfolio(R=ret_edhec, portfolio=pspec_maxret,
                                 optimize_method=c("CVXR", "SCS"), trace=TRUE,MaxSR=TRUE)
opt_maxret

bt_maxret <- optimize.portfolio.rebalancing(R=ret_edhec, portfolio=pspec_maxret,
                                            optimize_method=c("CVXR", "SCS"),
                                            rebalance_on="quarters",
                                            training_period=36, trace=TRUE, MaxSR=TRUE)