Chop DAS data into effort segments by continuous effort section
Usage
das_chop_section(x, ...)
# S3 method for class 'data.frame'
das_chop_section(x, ...)
# S3 method for class 'das_df'
das_chop_section(x, conditions, distance.method = NULL, num.cores = NULL, ...)
Arguments
- x
an object of class
das_df
, or a data frame that can be coerced to classdas_df
This data must be filtered for 'OnEffort' events; see the Details section below- ...
ignored
- conditions
see
das_effort
- distance.method
character; see
das_effort
. Default isNULL
since these distances should have already been calculated- num.cores
see
das_effort
Value
See das_chop_equallength
. The randpicks values will all be NA
Details
WARNING - do not call this function directly! It is exported for documentation purposes, but is intended for internal package use only.
This function is simply a wrapper for das_chop_equallength
.
It calls das_chop_equallength
, with seg.km
set to a
value larger than the longest continuous effort section in x
.
Thus, the effort is 'chopped' into the continuous effort sections and then summarized.
See the Examples section for an example where the two methods give the same output. Note that the longest continuous effort section in the sample data is ~22km.
Examples
y <- system.file("das_sample.das", package = "swfscDAS")
y.proc <- das_process(y)
y.eff1 <- das_effort(y.proc, method = "equallength", seg.km = 25, num.cores = 1)
#> No argument was passed via randpicks.load, and thus new randpicks values will be generated
y.eff2 <- das_effort(y.proc, method = "section", num.cores = 1)
all.equal(y.eff1, y.eff2)
#> [1] TRUE