bpaggregate: Apply a function on subsets of data frames
In BiocParallel: Bioconductor facilities for parallel evaluation

Description Usage Arguments Details Value Author(s) Examples

This is a parallel version of aggregate.

## S4 method for signature 'formula,BiocParallelParam'
bpaggregate(x, data, FUN, ..., 
    BPREDO=list(), BPPARAM=bpparam())

## S4 method for signature 'data.frame,BiocParallelParam'
bpaggregate(x, by, FUN, ..., 
    simplify=TRUE, BPREDO=list(), BPPARAM=bpparam())

## S4 method for signature 'matrix,BiocParallelParam'
bpaggregate(x, by, FUN, ..., 
    simplify=TRUE, BPREDO=list(), BPPARAM=bpparam())

## S4 method for signature 'ANY,missing'
bpaggregate(x, ..., BPREDO=list(), BPPARAM=bpparam())

`x`	A `data.frame`, `matrix` or a formula.
`by`	A list of factors by which `x` is split; applicable when `x` is `data.frame` or `matrix`.
`data`	A `data.frame`; applicable when `x` is a `formula`.
`FUN`	Function to apply.
`...`	Additional arguments for `FUN`.
`simplify`	If set to `TRUE`, the return values of `FUN` will be simplified using `simplify2array`.
`BPPARAM`	An optional `BiocParallelParam` instance determining the parallel back-end to be used during evaluation.
`BPREDO`	A `list` of output from `bpaggregate` with one or more failed elements. When a list is given in `BPREDO`, `bpok` is used to identify errors, tasks are rerun and inserted into the original results.

bpaggregate is a generic with methods for data.frame matrix and formula objects. x is divided into subsets according to factors in by. Data chunks are sent to the workers, FUN is applied and results are returned as a data.frame.

The function is similar in spirit to aggregate from the stats package but aggregate is not explicitly called. The bpaggregate formula method reformulates the call and dispatches to the data.frame method which in turn distributes data chunks to workers with bplapply.

See aggregate.

Martin Morgan mailto:mtmorgan@fhcrc.org.

if (interactive() && require(Rsamtools) && require(GenomicAlignments)) {

  fl <- system.file("extdata", "ex1.bam", package="Rsamtools")
  param <- ScanBamParam(what = c("flag", "mapq"))
  gal <- readGAlignments(fl, param=param) 

  ## Report the mean map quality by range cutoff:
  cutoff <- rep(0, length(gal))
  cutoff[start(gal) > 1000 & start(gal) < 1500] <- 1
  cutoff[start(gal) > 1500] <- 2 
  bpaggregate(as.data.frame(mcols(gal)$mapq), list(cutoff = cutoff), mean)

}