Grouping data from data frame

Question

I have a data that has following form:

first second data_col1 data_col2 data_col3
lu    NA     <number>  <number>  <number>
lu    NA     <number>  <number>  <number>
lu    NA     <number>  <number>  <number>
lu    NA     <number>  <number>  <number>
lu    mult   <number>  <number>  <number>
lu    mult   <number>  <number>  <number>
lu    mult   <number>  <number>  <number>
lu    mult   <number>  <number>  <number>
mult  NA     <number>  <number>  <number>
mult  NA     <number>  <number>  <number>
mult  NA     <number>  <number>  <number>
mult  NA     <number>  <number>  <number>

And so on.

I want to group this data by first two columns and plot them separately.

I try to do this:

comb <- unique(total.df[c(1,2)])
apply(comb, 1, function(x) {
  d<-total.df[total.df$guess==FALSE &
              total.df$second==x[2] &
              total.df$first==x[1] &
              total.df$tasks=='tasks_const',]
  p = ggplot(d, aes(x=d$platform, y=d$time,
                    group=as.factor(d$sched),
             colour=as.factor(d$sched))) +
      geom_point() + geom_line()
  ggsave(filename=sprintf("/tmp/a_%s_%s.png", x[1], x[2]))
})

My comb looks following:

        first   second
1        mult     <NA>
121        lu     mult
241        lu     <NA>
361      heat     mult
481      heat       lu
601      heat     <NA>
721  cholesky     mult
841  cholesky       lu
961  cholesky     heat
1081 cholesky     <NA>
1201 pipeline     mult
1321 pipeline       lu
1441 pipeline     heat
1561 pipeline cholesky
1681 pipeline     <NA>
1801      gen     mult
1921      gen       lu
2041      gen     heat
2161      gen cholesky
2281      gen pipeline
2401      gen     <NA>

facet_wrap almost solves my task, but I want every picture to be separate to able to see what is actually there. And with facet_wrap each of them is too small.

Code with facet_wrap is following:

ggplot(total.df, aes(x=total.df$platform, y=total.df$time,
       group=as.factor(total.df$sched),
       colour=as.factor(total.df$sched))) +
geom_point() + geom_line() + facet_wrap(first ~ second);

You don't need to specify total.df in your call of aes(). The first argument of ggplot specifies the data frame. So aes(x=total.df$platform,... can be simplified to aes(x = platform) for example. — davechilders
– davechilders, Commented Jan 27, 2015 at 21:40

Mahrtynas · Accepted Answer · 2015-01-28 08:03:59Z

2

I would probably suggest plotting each graph on a different page of one pdf file. I could also recommend using data.table, as it makes things look nicer:

library(data.table)
total.dt <- data.table(total.df)
setkey(total.dt, first, second)
comb <- unique(total.dt[, list(first, second)])

pdf("test.pdf")
for(n in 1:nrow(comb)){
  d <- total.dt[comb[n, ]][guess == FALSE & tasks == "tasks_const"] 
  print(ggplot(d, aes(x = platform, y = time,
                      group = as.factor(sched),
                      colour = as.factor(sched))) +
        geom_point() + geom_line() + 
        ggtitle(sprintf("first=%s, second=%s", comb[n, first], comb[n, second])))}
dev.off()

answered Jan 28, 2015 at 8:03

Mahrtynas

1062 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Grouping data from data frame

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related