tests/limitations/test_rank.R

In Fuzzy-Logix/AdapteR: Converts R syntax to SQL to provide fast analytics using in-database C++ library DB Lytix.

Renv = new.env(parent = globalenv())

Renv$var1 = c(3, 1, 4, 15, 92)
Renv$var2 = c(3, 1, 4, 1, 5, 9, 2, 6, 5, 3, 5)
names(Renv$var2) = letters[1:11]
Renv$vector1 = sample(1:100,60,replace = TRUE)

FLenv = as.FL(Renv)

## Test2 fails because indexColumn is lost after function is applied,
## so the order of vector is lost. Observed in case of duplicate values.
## https://app.asana.com/0/150173007236461/233442378511381
test_that("rank: default method average, named vectors",{
    result = eval_expect_equal({
        test1 = rank(var1)
        test2 = rank(var2)
    },Renv,FLenv,
    expectation=c("test1","test2"))
    ##print(result)
})

test_that("rank: check for idempotent characteristic of rank function with ties.method=perc, duplicate",{
    result = eval_expect_equal({
        rtest1 <- rank(test1)
        expect_equal(as.vector(rtest1),as.vector(test1))
        rtest2 <- rank(test2)
        expect_equal(as.vector(rtest2),as.vector(test2))
    },Renv,FLenv)
    ##print(result)
})


#AdapteR supports perc and frac methods other than average method for tie breaking.
#SQL Query takes lot of time.
test_that("rank(FLVector): ",{
    test3 = rank(FLenv$vector1,ties.method = "perc")
    expect_equal(as.vector(rank(test3,ties.method = "perc")),
                 as.vector(test3))
    test4 = rank(FLenv$vector1,ties.method = "duplicate")
    expect_equal(as.vector(rank(test4,ties.method = "duplicate")),
                 as.vector(test4))
})

## All these methods are not supported.. DB-Lytix and  R have average.
## In addition FL has "perc" and "duplicate" which are not in R
for(meth in c("first","last","random","max","min"))
    tryCatch(
        test_that(paste0("rank: base R ties.method=",meth),{
            result = eval_expect_equal({
                testm = rank(var2, ties.method= meth)  
            },Renv,FLenv,
            expectation="testm")
        }),
        error=function(e) e)