How to compute maximum or minimum value between two dates that are not hardcoded

[BitRacketeer]

I've been grappling with this one put can't crack it.  Have checked the online forums, xts docs etc. but seems this is one step beyond what's there.

What I'm doing
1. start with a time-series
2. filter the time-series in 1 in some way
3. filter the results in 2 in some way, returning pairs of dates

What I then want to do: screen the original time-series in step 1 for the Max and Min price between the pair of dates in Step 3.

I can filter an xts in a hard-coded way e.g.

xts["2012-08-01::2012-08-05"]

but I've tried with a variable and it doesn't work.  It just passes everything through:

originalTimeSeries[paste(as.Date(combined[[1]]), "::", as.Date(combined[[2]]))]

where combined is

     firstDate secondDate

[1,]     -2747      -1179

[2,]      1737       2981

[3,]      2981       3738

[4,]      6501       7588

[5,]      8470       8796

(for whatever reasons the dates come through as longs.)


and created from
firstDate <- <some filter from the time series>

secondDate <- <some filter from the time series>

combined <- cbind(firstDate,secondDate)


So what I want is some kind of way to go
    max(originalTimeSeries[between firstDate and secondDate])
where firstDate and secondDate are the results of some other query

for each line in combined.  Effectively this is just a min-max exercise where the window-size is determined by some other criteria, dynamically, and will typically change with each run.

Any ideas on how to get the results all chained together?  Clearly it can be done with hard-coded dates but that defeats the point of the exercise.

[BitRacketeer]

I think I've got there ... the trick is that paste() introduces a separator.

This is solved by
originalTimeSeries[paste(as.Date(combined[[1]]), "::", as.Date(combined[[2]]),sep="")]


Is there a more elegant way however?