Importing Rds R model to Power BI - gam model

[patipati]

Hi,

I'm trying to pass parameters of Gam model using this method.

https://towardsdatascience.com/how-to-predict-values-from-a-custom-r-model-in-power-bi-3364f83b0015

Unfortunettaly I am gettign read error on unserialize function

from_byte_string = function(x) {

xcharvec = strsplit(x, " ")[[1]]

xhex = as.hexmode(xcharvec)

xraw = as.raw(xhex)

unserialize(xraw)

 

}

#project_folder <- "Z:/Datasets/NSA/EBITDA_adimensional/"

#finalfit <- readRDS(paste0(project_folder, "mtcars-model.rds"))

finalfit.str <- as.character( dataset[dataset$model_id == 1, "model_str"] )

finalfit <- from_byte_string(finalfit.str)

v <- c( wt=dataset$weight, qsec=dataset$qsec, am=as.factor(dataset$am) )

pred <- round( predict(finalfit, newdata = as.data.frame(t(v))), digits = 2 )

plot.new()

text(0.5, 0.5,

#labels = as.character(dataset$weight)

labels = as.character(pred[[1]])

,cex = 3.5

)

On IDE script is runnig properly.

Is there anyone who had tried it yet ? Can it be problem with library in Power BI ?

[v-chuncz-msft]

@patipati 

That works for me. Try the latest version of Power BI Desktop and take a good look at Create Power BI visuals with R.

[patipati]

I have March version, and R 3.5.3 and I'm still getting error with unserialize function:

unserialize(xraw): read error

[Anonymous]

Where you able to solve this issue?