In real-world analytics, you often need to integrate information from multiple sources. This means joining several data frames and making decisions about how to handle missing values that result from incomplete matches across those sources. You will now apply your joining and data cleaning skills to build a unified dataset that could power business analysis.


library(testthat)

source("user_code.R")
test_that("all customers included in result", {
  res <- integrate_data_sources(customers, orders, payments)
  # All customer IDs from customers should appear at least once, or as NA if only in orders
  cust_ids <- unique(customers$customer_id)
  expect_true(all(cust_ids %in% c(res$customer_id, NA)),
    info = "All customer IDs from customers should be present in the result.")
})

test_that("all orders included in result", {
  res <- integrate_data_sources(customers, orders, payments)
  order_ids <- unique(orders$order_id)
  expect_true(all(order_ids %in% res$order_id),
    info = "All order IDs from orders should be present in the result.")
})

test_that("all payments included in result", {
  res <- integrate_data_sources(customers, orders, payments)
  payment_ids <- unique(payments$payment_id)
  expect_true(all(payment_ids %in% c(res$payment_id, NA)),
    info = "All payment IDs from payments should be present in the result (NA allowed for unmatched).")
})

test_that("missing names are replaced with 'Unknown'", {
  res <- integrate_data_sources(customers, orders, payments)
  expect_false(any(is.na(res$name)),
    info = "There should be no NA in the name column.")
  expect_true(any(res$name == "Unknown"),
    info = "There should be at least one 'Unknown' in the name column where customer is missing.")
})

test_that("missing order_total and payment_amount are zero", {
  res <- integrate_data_sources(customers, orders, payments)
  expect_false(any(is.na(res$order_total)),
    info = "There should be no NA in order_total column.")
  expect_false(any(is.na(res$payment_amount)),
    info = "There should be no NA in payment_amount column.")
  expect_true(all(res$order_total >= 0),
    info = "All order_total values should be zero or positive.")
  expect_true(all(res$payment_amount >= 0),
    info = "All payment_amount values should be zero or positive.")
})

test_that("correct joining on customer_id and order_id", {
  res <- integrate_data_sources(customers, orders, payments)
  # Check that for a row with customer_id==1 and order_id==100, name is Alice, order_total is 250, payment_amount is 250
  row <- res[res$customer_id == 1 & res$order_id == 100, ]
  expect_true(nrow(row) == 1, info = "There should be exactly one row for customer_id 1 and order_id 100.")
  expect_equal(row$name, "Alice", info = "Name should be Alice for customer_id 1.")
  expect_equal(row$order_total, 250, info = "Order total should be 250 for order_id 100.")
  expect_equal(row$payment_amount, 250, info = "Payment amount should be 250 for order_id 100.")
})

test_main.R

Master the essential data manipulation skills in R using dplyr, grouping and aggregation, and joining data frames. This course is designed for beginners and provides hands-on, real-world tasks to build your confidence in transforming and analyzing data for analytics.

Learn the foundational verbs of dplyr for selecting, filtering, mutating, arranging, and summarizing data. Build confidence in transforming data frames for analytics.

Master the use of pipes (%>%) to chain multiple data manipulation steps for readable and efficient R code.

Learn to group data and calculate aggregated metrics for powerful reporting and analytics.

Master the art of combining data from multiple sources using different types of joins in dplyr.

Challenge: Integrate Multiple Data Sources

Lösning

Challenge: Integrate Multiple Data Sources

Lösning