Issue on cointegration with only one dumvar, urca

I am using urca package to get Johansen test results for 3 variables with 2 dummy variables:

library(urca)

johansenresults <- ca.jo(datamod[,c(2,11,9)], ecdet = "const", type="trace", K=5, spec="longrun",season=NULL, dumvar = datamod2[,c(13,14)])

Everything works fine, and all later estimation related to cajorls and vec2var also work. However, when I want to estimate the same model with only one dummy variable I get this error:

nplnbelcpivecmtrace2 <- ca.jo(datamod[,c(2,11,9)], ecdet = "const", type="trace", K=5, spec="longrun",season=NULL, dumvar = datamod2[,13])

Warning message:
In ca.jo(datamod[, c(2, 11, 9)], ecdet = "const", type = "trace",  : 
No column names in 'dumvar', using prefix 'exo' instead.

Does anyone have a similar experience?