tidync (version 0.4.0)

hyper_filter: Subset NetCDF variable by expression

Description

The hyper_filter() acts on a tidync object by matching one or more filtering expressions like with dplyr::filter. This allows us to lazily specify a subset from a NetCDF array without pulling any data. The modified object may be printed to see the effects of subsetting, or saved for further use.

Usage

hyper_filter(.x, ...)
# S3 method for tidync
hyper_filter(.x, ...)

Value

data frame

Arguments

.x

NetCDF file, connection object, or tidync object

...

currently ignored

Details

The function hyper_filter() will act on an existing tidync object or a source string.

Filter arguments must be named as per the dimensions in the variable in form dimname = dimname < 10. This is a restrictive variant of dplyr::filter(), with a syntax more like dplyr::mutate(). This ensures that each element is named, so we know which dimension to apply this to, but also that the expression evaluated against can do some extra work for a nuanced test.

There are special columns provided with each axis, one is 'index' so that exact matching can be done by position, or to ignore the actual value of the coordinate. That means we can use a form like dimname = index < 10 to subset by position in the array index, without necessarily knowing the values along that dimension.

Examples

f <- "S20080012008031.L3m_MO_CHL_chlor_a_9km.nc"
l3file <- system.file("extdata/oceandata", f, package= "tidync")
## filter by value
tidync(l3file) %>% hyper_filter(lon = lon < 100)
## filter by index
tidync(l3file) %>% hyper_filter(lon = index < 100)

## be careful that multiple comparisons must occur in one expression
 tidync(l3file) %>% hyper_filter(lon = lon < 100 & lon > 50)

## filter in combination/s
tidync(l3file) %>% hyper_filter(lat = abs(lat) < 10, lon = index < 100)