demonstrate use of dplyr::filter with lapply, both with and without NSE

example_filter_with_nse.R

# Example R code to filter data using criteria passed in via lists, using lapply
# Matt Jones 2015-10-19

library(dplyr)
# create a sample fake data frame
df <- cbind(expand.grid(sciName=list("A", "B", "C", "D"), family=list("X", "Y"), stage=list("S1", "S2", "S3", "S4")), count=1)
df

# set up our filter conditions
condition1 <- list(sciName="A", stageVector=c("S2", "S3"))
condition2 <- list(sciName="C", stageVector=c("S3", "S4"))
conditionList <- list(condition1, condition2)

# Define the filtering function
filterStages <- function(condition, df) {
    subset.data <- df %>%
        filter(sciName == condition$sciName) %>%
        filter(stage %in% condition$stageVector)
    return(subset.data)
}

# demo the filter function working on a single condition at a time
filterStages(condition1, df)
filterStages(condition2, df)

# demo the filter function working over a list of conditions
resultDataList <- lapply(conditionList, filterStages, df)
resultDataList


## TRY IT ALL DIFFERENTLY WITH SE rather than NSE so that we can choose filter variables dynamically
library(lazyeval)
condition1 <- list(taxonField="sciName", taxonValue="A", stageVector=c("S2", "S3"))
condition2 <- list(taxonField="family", taxonValue="Y", stageVector=c("S3", "S4"))
conditionList <- list(condition1, condition2)

filterStages <- function(condition, df) {
    cond <- interp(~ tf == tv, tf = lazy(condition$taxonField), tv = condition$taxonValue)
    subset.data <- df %>%
        filter_(interp(~ tf == condition$taxonValue, tf = as.name(condition$taxonField))) %>%
        filter(stage %in% condition$stageVector)
    return(subset.data)
}

filterStages(condition1, df)
filterStages(condition2, df)

resultDataList <- lapply(conditionList, filterStages, df)
resultDataList