tests/testthat/test-merge.R

In jackwasey/jwutil: Tools for Data Manipulation and Testing

context("test merging")

dfa <- dfb <- dfc <- data.frame(
  a = c(1, 2, 3, 4),
  b = c(11, 12, 13, 14),
  c = c(101, 102, 103, 104)
)
dfb[1, "b"] <- 999
dfc[1, "b"] <- 999
dfc[2, "c"] <- 888

test_that("merge identical frames give identical result unless affix", {
  r <- mergeBetter(x = dfa, y = dfa, by.x = "a", by.y = "a", verbose = FALSE)
  expect_equal(dfa, r)
})

test_that("identical frames with reordered should give identical result", {
  expect_equal(mergeBetter(
    x = dfa, y = dfa[c("a", "c", "b")],
    by.x = "a", by.y = "a"
  ), dfa)
  expect_equal(mergeBetter(
    x = dfa, y = dfa[c("c", "b", "a")],
    by.x = "a", by.y = "a"
  ), dfa)
})

test_that("can't handle double duplicate fields, esp not keys", {
  df_a <- dfa[c("a", "b", "c", "c")]
  names(df_a) <- c("a", "b", "C", "c")
  expect_error(mergeBetter(x = dfa, y = df_a, by.x = "a", by.y = "a"))
  df_a <- dfa[c("a", "b", "b", "c")]
  names(df_a) <- c("a", "b", "B", "c")
  expect_error(mergeBetter(x = dfa, y = df_a, by.x = "a", by.y = "a"))
  df_a <- dfa[c("a", "a", "b", "c")]
  names(df_a) <- c("a", "A", "b", "c")
  expect_error(mergeBetter(x = dfa, y = df_a, by.x = "a", by.y = "a"))
  df_a <- dfa[c("a", "a", "b", "b")]
  names(df_a) <- c("a", "A", "B", "b")
  expect_error(mergeBetter(x = dfa, y = df_a, by.x = "a", by.y = "a"))
})

test_that("merge non-identical frames should suffix field name by default", {
  r <- mergeBetter(x = dfa, y = dfb, by.x = "a", by.y = "a", verbose = FALSE)
  e <- dfa
  e[["b.dfb"]] <- c(999, 12, 13, 14)
  expect_equal(r, e)
})

test_that("merge identical frames should give identical result", {
  r <- mergeBetter(x = dfa, y = dfc, by.x = "a", by.y = "a")
  e <- dfa
  e[["b.dfc"]] <- c(999, 12, 13, 14)
  e[["c.dfc"]] <- c(101, 888, 103, 104)
  expect_identical(r, e)
})

test_that("merge with missing id field fails with error", {
  expect_that(
    mergeBetter(x = dfa, y = dfb, by.x = "sloe", by.y = "gin"),
    throws_error()
  )
})

test_that("merge frames forcing prefix", {
  r <- mergeBetter(
    x = dfa, y = dfa, by.x = "a", by.y = "a",
    renameAll = "prefix"
  )

  r <- mergeBetter(
    x = dfa, y = dfc, by.x = "a", by.y = "a",
    renameAll = "prefix"
  )
  e <- structure(list(
    a = c(1, 2, 3, 4),
    b = c(11, 12, 13, 14),
    c = c(101, 102, 103, 104),
    dfc.b = c(999, 12, 13, 14),
    dfc.c = c(101, 888, 103, 104)
  ),
  .Names = c("a", "b", "c", "dfc.b", "dfc.c"),
  class = "data.frame"
  )
  expect_equivalent(r, e)

  r <- mergeBetter(
    x = dfa, y = dfc, by.x = "a", by.y = "a",
    renameAll = "suffix", verbose = FALSE
  )
  e <- structure(list(
    a = c(1, 2, 3, 4),
    b = c(11, 12, 13, 14),
    c = c(101, 102, 103, 104),
    dfc.b = c(999, 12, 13, 14),
    dfc.c = c(101, 888, 103, 104)
  ),
  .Names = c("a", "b", "c", "b.dfc", "c.dfc"),
  class = "data.frame"
  )
  expect_equivalent(r, e)

  r <- mergeBetter(
    x = dfa, y = dfc, by.x = "a", by.y = "a",
    renameAll = "suffix", affix = "jack"
  )
  e <- structure(list(
    a = c(1, 2, 3, 4),
    b = c(11, 12, 13, 14),
    c = c(101, 102, 103, 104),
    dfc.b = c(999, 12, 13, 14),
    dfc.c = c(101, 888, 103, 104)
  ),
  .Names = c("a", "b", "c", "b.jack", "c.jack"),
  class = "data.frame"
  )
  expect_equivalent(r, e)

  # should be case insensitive for field matches (but not the 'by' term)
  df_b <- dfb
  names(df_b) <- c("a", "B", "C")
  r <- mergeBetter(
    x = dfa, y = df_b, by.x = "a", by.y = "a",
    renameAll = "suffix", affix = "jack"
  )
  e <- structure(list(
    a = c(1, 2, 3, 4),
    b = c(11, 12, 13, 14),
    c = c(101, 102, 103, 104),
    dfc.b = c(999, 12, 13, 14),
    dfc.c = c(101, 102, 103, 104)
  ),
  .Names = c("a", "b", "c", "B.jack"),
  class = "data.frame"
  )
  expect_equivalent(r, e)

  skip("to do tests covering case where there are duplicate
       frames but rename is explicitly requested")
})