DImodelsVis (version 1.0.1)

custom_filter: Special custom filtering for compositional data

Description

A handy wrapper around the dplyr filter() function enabling the user to filter rows which satisfy specific conditions for compositional data like all equi-proportional communities, or communities with a given value of richness without having to make any changes to the data or adding any additional columns. All other functionalities are same as the dplyr filter() function.

Usage

custom_filter(data, prop = NULL, special = NULL, ...)

Value

A subset of the original data which matches the specified filtering conditions.

Arguments

data

A data frame containing the compositional variables which should be used to perform the filtering.

prop

A character/numeric vector indicating the columns containing the compositional variables in `data`.

special

A character string specifying the filtering condition. Four special keywords can be specified here for filtering 1. richness: A positive integer value to filter communities with a specific number of compositional variables (variables with non-zero values). 2. evenness: A numeric value between 0 and 1, to filter rows based on the relative abundances of the compositional variables where a higher value signifies a more even community with equal proportions of all variables. 3. equi: A boolean variable indicating whether to filter rows containing equi-proportional communities, i.e., communities where all variables have the same non-zero proportion. 4. monos: A boolean value indicating whether to filter communities containing a single compositional variable, i.e., richness == 1. These keywords can be combined using any logical operators and can even be combined with any other variables in the data. Please use the exact keywords (case-sensitive) in the query to get appropriate results. See examples for more details.

...

Any additional arguments specified to the dplyr filter() function. Filtering conditions for any additional variables can also be specified here.

Examples

library(DImodels)
library(dplyr)

## Load data
data(sim3)

# The special filter keywords should be specified as a string
# Filter communities containing 3 species
head(custom_filter(data = sim3, prop = 4:12,
                   special = "richness == 3"))

# Filter communities at richness 6 OR evenness 0
head(custom_filter(data = sim3, prop = 4:12,
                   special = "richness == 6 | evenness == 0"), 12)

# Filter all monoculture AND treatment "A" (treatment is column present in data)
head(custom_filter(data = sim3, prop = 4:12,
                   special = "monos == TRUE & treatment == 'A'"), 10)

# Filter all equi proportional communities but NOT monocultures
head(custom_filter(data = sim3, prop = 4:12,
                   special = "equi == TRUE & monos == FALSE"))

# Can also use normal filter
sim3 %>% custom_filter(p1 == 1, special = NULL, prop = NULL)

# Both special filtering and normal filtering can be combined as well
sim3 %>% custom_filter(prop = paste0("p", 1:9),
                       special = "richness == 1",
                       community %in% c(7, 9))