ggplot-How to create a legend using row&column names - r

Question

I have data frame which I want to pass first two columns rows and
variable column names to create legend.
Inside of df I have group of dataset in which they grouped with letters from a to h. In particular, I want to pass AC&AR columns rows as names in combination with DQ0:DQ2 variables and they should be shown in the legend with that format.
something like 78_256_DQ0, and 78_256_DQ1 and 78_256_DQ2 for data group a
and same for the rest of letters in the df.
my reproducible df like this;
df <- do.call(rbind,lapply(1,function(x){
AC <- as.character(rep(rep(c(78,110),each=10),times=3))
AR <- as.character(rep(rep(c(256,320,384),each=20),times=1))
V <- rep(c(seq(2,40,length.out=5),seq(-2,-40,length.out=5)),times=2)
DQ0 = sort(replicate(6, runif(10,0.001:1)))
DQ1 = sort(replicate(6, runif(10,0.001:1)))
DQ2 = sort(replicate(6, runif(10,0.001:1)))
No = c(replicate(1,rep(letters[1:6],each=10)))
data.frame(AC,AR,V,DQ0,DQ1,DQ2,No)
}))
head(df)
AC AR V DQ0 DQ1 DQ2 No
1 78 256 2.0 0.003944916 0.00902776 0.00228837 a
2 78 256 11.5 0.006629239 0.01739512 0.01649540 a
3 78 256 21.0 0.048515226 0.02034436 0.04525160 a
4 78 256 30.5 0.079483625 0.04346118 0.04778420 a
5 78 256 40.0 0.099462310 0.04430493 0.05086738 a
6 78 256 -2.0 0.103686255 0.04440260 0.09931459 a
*****************************************************
library(reshape2)
df_new <- melt(df,id=c("V","No"),measure=c("DQ0","DQ1","DQ2"))
library(ggplot2)
ggplot(df_new,aes(y=value,x=V,group=No,colour=No))+
geom_point()+
geom_line()
UPDATE
after #... answer I made a little bit progress. His solution is partially ok. Because when we melt names
df$names <- interaction(df$AC,df$AR,names(df)[4:6])
df_new <- melt(df,id=c("V","No","names1"),measure=c("DQ0","DQ1","DQ2"))
this command plots 4 rows for each group a to h.
the output becomes like this;
head(df)
AC AR V DQ0 DQ1 DQ2 No names
1 78 256 2.0 0.002576547 0.04294134 0.008302918 a 78.256.DQ0
2 78 256 11.5 0.010150299 0.04570650 0.011749370 a 78.256.DQ1
3 78 256 21.0 0.012540026 0.06977744 0.013887357 a 78.256.DQ2
4 78 256 30.5 0.036532977 0.11460343 0.071172301 a 78.256.DQ0
5 78 256 40.0 0.042801967 0.11518191 0.073756228 a 78.256.DQ1
6 78 256 -2.0 0.043275144 0.13033194 0.076569977 a 78.256.DQ2
**************************************************************
and with modification of the plot command
ggplot(df_new,aes(y=value,x=V,lty=variable,colour=names))+
geom_point()+
geom_line()
the output format which I prefer is something I can refer all rows of DQ0,DQ1 and DQ2 inside of each group. Any suggestions?
last condition

Answer 1

u can use df$names <- interaction(v$AC,v$AR,DQ0) and then also set names in you melt command as id. Later you use color=names in your aes function.
So, this will add a column name with a combination of the defined columns. You can also set a sep='_' if you prefer over ..
If you now use this column for colouring, you will get those labels as legend names.

Answer 2

finally I found a way using gather from dplyr.
df_gather <- df %>% gather(DQ, value,-No, -AC, -AR, -V)
and using interaction function from #drmariod answer
df_gather$names <- interaction(df_gather$AC,df_gather$AR,df_gather$DQ)
and here is the result of this question:)