Question

I have a csv that I have imported that has multiple data fields for a times series. So the first field is the datetime, and the rest of the fields are variuous datapoints.

How would I plot the pointwise max value of multiple fields, where the fields would be matched on some sort of wildcard?

So for example:

time,foo1,foo2,foo3
1:00,1,2,3
2:00,3,1,1
3:00,2,5,3

What would be the easiest way to plot foo.* such that I get the max from each field: I.e. the resultant plot from that example would be: (1:00,3),(2:00,3),(3:00,5)?

To Clarify the example, the max points I mean are in **

time,foo1,foo2,foo3
1:00,1,2,*3*
2:00,*3*,1,1
3:00,2,*5*,3

Answer 1

Assuming your data are in a data.frame x, you can use pmax like so:

cbind(x[,"time"], do.call(pmax, x[,grepl("^foo*",names(x))]))

do.call calls the pmax function with each column in x as a ... argument (except the first column, which is removed by negative subscripting). cbind combines the first column of x with the resulting vector from do.call.

Note that the 2nd argument to do.call needs to be a list and a data.frame is a list with some extra attributes. If x isn’t a data.frame, you’ll need to coerce it to one (or a regular list).