Can't set conditional formating for colors in clustered column chart

[Anonymous]

If I have only one series of data I can chose a formula for the bar colors under "data colors". As soon as I add another data series to the chart this option is gone and I only can select a fixed color.

Is there no way to set the colors of multiple series by the value of fields?

Thanks!

[Icey]

Hi @Anonymous ,

Try to create a R visual.

library(ggplot2)
data<-dataset
data<-with(data,data[order(Category,ActualBudget),])
ggplot(data=data, aes(x=ActualBudget, y=Value,fill = ActualBudget)) 
       + geom_bar(stat="identity")
       +facet_grid(~Month)
       +scale_fill_manual(values=c("red","blue"))  -------- set the color you want

Best regards

Icey

If this post helps,then consider Accepting it as the solution to help other members find it faster.

[Anonymous]

Hi Icey

thanks, this is a good idea.

Still I can't build the chart I need.

Here the visual as it should look like:

Here the data handed over to the visual:

And what I'm able to create:

Here my code:

library(ggplot2)
theme_set(theme_classic())

dataCleaned <- subset(dataset, !is.na(SupplierAvail100))
dataCleaned$Date <- as.Date(dataCleaned$Date)
dataCleaned$SupplierAvail100 <- round(dataCleaned$SupplierAvail100, digits = 1)

labl = paste0(month.abb[strtoi(substr(dataCleaned$Date,6,7),10)], " ", substr(dataCleaned$Date,1,4))
brks = dataCleaned$Date

ggplot(data=dataCleaned, aes(x=Date))+
  geom_col(aes(y=OTIF100), show.legend = FALSE, position = "dodge") +
  geom_col(aes(y=SupplierAvail100), show.legend = FALSE, fill = dataCleaned$SupplierAvailColor, position = "dodge") +
  geom_text(aes(x = Date, y = SupplierAvail100, label = SupplierAvail100, vjust = -0.5)) +
  theme(axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12))+
  scale_x_continuous(labels = labl, breaks = brks, guide = guide_axis(check.overlap = TRUE))+
  theme(line = element_blank(), axis.title.x = element_blank(), axis.title.y = element_blank())

 I don't know if a such complex visual is doable.

I'm very interested in your opinion.

Thanks

Rolf

[Icey]

Hi @Anonymous ,

Could you share some smaple data with the same structure of your real data for test? And please remove sensitive information.

Best regards

Icey

[Anonymous]

Hi Icey

find attached a small demo of the problem. Everything works except that can have the colored bars (ColA) side-by-side with the other bar (ColC).

R Script demo 

Do you have me a hint?

Thanks!

[amitchandak]

@Anonymous , You can not use conditional formatting with more than one measure or legend as of now.

You can vote for an idea or log a new one -https://ideas.powerbi.com/ideas/

[Anonymous]

Thanks for the fast answer.

Is is then at least possible to have one serie with a measure and all others with static colors?