Six rows from Cartwheel’s messy export sit in a list of dictionaries. You want their total value, how much came through each sales channel, and which orders were large. Writing one line per row would work for six rows and collapse at six thousand. A loop writes the logic once and Python repeats it for every row.
\norders = [\n {\"order_id\": 2574, \"total\": 65.0, \"channel\": \"mobile app\"},\n {\"order_id\": 2069, \"total\": 18.0, \"channel\": \"mobile app\"},\n {\"order_id\": 1200, \"total\": 33.0, \"channel\": \"mobile app\"},\n {\"order_id\": 1809, \"total\": 35.0, \"channel\": \"web\"},\n {\"order_id\": 1434, \"total\": 72.0, \"channel\": None},\n {\"order_id\": 2089, \"total\": 58.0, \"channel\": \"marketplace\"},\n]\n\ngrand_total = 0\nfor order in orders:\n grand_total += order[\"total\"]\nprint(grand_total)Read it as “for each order in orders, run the indented lines”. On each pass, the name order points at the next dictionary. grand_total += order["total"] is short for grand_total = grand_total + order["total"].
Indentation is not decoration in Python: the indented lines (four spaces, by convention) are the loop’s body, and the first line back at the left margin runs after the loop finishes. That is why print sits outside: you want one result, not six.
Conditions let each row take a different path. Cartwheel’s operations team labels orders by size:
\ntotals = [65.0, 18.0, 33.0, 35.0, 72.0, 58.0, 329.0]\n\nfor t in totals:\n if t < 50:\n size = \"small\"\n elif t < 150:\n size = \"medium\"\n else:\n size = \"large\"\n print(t, size)Python tests the conditions from top to bottom and runs only the first branch that is true. Order matters: if the 150 test came first, every small order would be labelled medium too, because 18 is also below 150. Put the narrowest condition first, or make the conditions non-overlapping.
\nYou can combine conditions with and, or and not. Python also lets you chain comparisons the way you would write them in maths, which is handy for ranges:
t = 89.0\nchannel = \"web\"\nprint(t >= 50 and channel == \"web\") # both must be true\nprint(50 <= t < 150) # chained: at least 50 and below 150\nprint(not channel == \"marketplace\") # flips True and FalseWhen you also need each item’s position, for a ranked list or a “row 3 is broken” message, wrap the list in enumerate(), which hands you the position and the item together: for position, t in enumerate(totals, start=1):. Prefer it to counting positions by hand with a separate variable.
The single most useful loop pattern in analysis is “per category, add something up”. A dictionary holds one running total per key:
\norders = [\n {\"order_id\": 2574, \"total\": 65.0, \"channel\": \"mobile app\"},\n {\"order_id\": 2069, \"total\": 18.0, \"channel\": \"mobile app\"},\n {\"order_id\": 1200, \"total\": 33.0, \"channel\": \"mobile app\"},\n {\"order_id\": 1809, \"total\": 35.0, \"channel\": \"web\"},\n {\"order_id\": 1434, \"total\": 72.0, \"channel\": None},\n {\"order_id\": 2089, \"total\": 58.0, \"channel\": \"marketplace\"},\n]\n\nrevenue_by_channel = {}\nfor order in orders:\n channel = order[\"channel\"] or \"unknown\"\n revenue_by_channel[channel] = revenue_by_channel.get(channel, 0) + order[\"total\"]\n\nfor channel, revenue in revenue_by_channel.items():\n print(channel, revenue)Two details carry this. revenue_by_channel.get(channel, 0) returns the running total so far, or 0 the first time a channel appears. And order["channel"] or "unknown" replaces None with a readable label, because None counts as false and or returns the right-hand side when the left is false.
This is a hand-written groupby. In Lesson 4.1 pandas does it in one line, and now you know what that line does underneath: walk the rows, find each row’s group, update that group’s total.
Very often a loop exists only to build a new list. A list comprehension says that directly:
\ntotals = [65.0, 18.0, 33.0, 35.0, 72.0, 58.0, 329.0]\n\nwith_rise = [round(t * 1.05, 2) for t in totals]\nlarge = [t for t in totals if t >= 60]\nprint(with_rise)\nprint(large, len(large))Read [t for t in totals if t >= 60] as “t, for each t in totals, if t is at least 60”. The part before for is what goes in; the if at the end filters. A dictionary comprehension works the same way with a key: value pair:
countries = [\"Canada\", \"united kingdom\", \"UNITED KINGDOM\", \"Saudi Arabia\"]\nclean = [c.strip().title() for c in countries]\nprint(clean)\nprint(sorted(set(clean))) # unique values, sorted\nprint({c: len(c) for c in set(clean)})A set keeps only unique values, so set(clean) collapses the two spellings of the UK into one once they are cleaned. You will do the same fix on a whole column with pandas in Lesson 3.2.
In pandas you will almost never loop over rows. orders["total"].sum() does the summing loop, and orders.groupby("channel")["total"].sum() does the dictionary pattern above, both in fast compiled code. So why learn loops? Because they are how you check your understanding of what a pandas operation does, and because plenty of real work (looping over files, over months, over a list of countries to build a report) is still a loop around pandas code.
In the exercise you summarise twelve rows by country. Then you package logic like this into reusable functions and learn to read the errors Python shows you.
\n","toc":[{"id":"sec-1","text":"The for loop","level":2},{"id":"sec-2","text":"Making decisions with if, elif, else","level":2},{"id":"sec-3","text":"Counting and summing by group","level":2},{"id":"sec-4","text":"Comprehensions: building lists in one line","level":2},{"id":"sec-5","text":"Loops versus pandas","level":2}],"takeaways":["Afor loop runs its indented body once per item; create accumulators such as total = 0 before the loop, not inside it.","if / elif / else checks conditions top to bottom and runs only the first branch that is true.","counts[key] = counts.get(key, 0) + 1 is the standard pattern for counting or summing by group with a dictionary.","A list comprehension [expr for x in items if condition] builds a new list in one readable line.","In pandas you rarely write loops, because column operations do the looping for you; loops are still how you reason about what those operations do."],"further":[{"title":"More Control Flow Tools (if, for, range)","url":"https://docs.python.org/3/tutorial/controlflow.html"},{"title":"List Comprehensions","url":"https://docs.python.org/3/tutorial/datastructures.html#list-comprehensions"}],"quiz":[{"q":"What does this print?
\ntotals = [65, 18, 33]\nfor t in totals:\n running = 0\n running += t\nprint(running)116","why":"That needs running = 0 before the loop. Here it is reset on every pass."},{"text":"33","why":"Correct. The accumulator is reset to 0 at the start of every pass, so only the last value survives."},{"text":"0","why":"The reset happens before the addition in each pass, so the final pass still adds 33."},{"text":"A NameError","why":"running is created inside the loop on the first pass, so it exists when print runs."}],"answer":1},{"q":"An order total of 150 goes through if t < 50: "small", elif t < 150: "medium", else: "large". Which label does it get?
"medium"","why":"150 < 150 is False, so the elif branch does not run."},{"text":""large"","why":"Correct. Both conditions are false for 150, so the else branch runs."},{"text":"Both "medium" and "large"","why":"An if/elif/else chain runs at most one branch, the first whose condition is true."}],"answer":1},{"q":"Which comprehension keeps only web orders’ totals from a list of order dictionaries?
","options":[{"text":"[o["total"] for o in orders if o["channel"] == "web"]","why":"Correct. The expression before for is what goes into the new list; the if at the end filters which items get there."},{"text":"[o for o in orders if o["channel"] == "web"]["total"]","why":"This builds a list of dictionaries and then indexes the list with a string, which raises TypeError."},{"text":"[o["total"] if o["channel"] == "web" for o in orders]","why":"A trailing filter goes after the for. An if before the for must have an else, so this is a syntax error."}],"answer":0}],"exercise":{"kind":"python","files":[],"starter":"orders = [\n {\"order_id\": 2574, \"total\": 65.0, \"channel\": \"mobile app\", \"country\": \"Canada\"},\n {\"order_id\": 2069, \"total\": 18.0, \"channel\": \"mobile app\", \"country\": \"United States\"},\n {\"order_id\": 1200, \"total\": 33.0, \"channel\": \"mobile app\", \"country\": \"united kingdom\"},\n {\"order_id\": 1809, \"total\": 35.0, \"channel\": \"web\", \"country\": \"United States\"},\n {\"order_id\": 1434, \"total\": 72.0, \"channel\": None, \"country\": \"Saudi Arabia\"},\n {\"order_id\": 2089, \"total\": 58.0, \"channel\": \"marketplace\", \"country\": \"UNITED KINGDOM\"},\n {\"order_id\": 2490, \"total\": 89.0, \"channel\": \"web\", \"country\": \"United Arab Emirates\"},\n {\"order_id\": 2985, \"total\": 24.0, \"channel\": \"web\", \"country\": \"United Kingdom\"},\n {\"order_id\": 2544, \"total\": 58.0, \"channel\": \"mobile app\", \"country\": \"Jordan\"},\n {\"order_id\": 2375, \"total\": 329.0, \"channel\": \"mobile app\", \"country\": \"United Kingdom\"},\n {\"order_id\": 1129, \"total\": 29.0, \"channel\": \"web\", \"country\": \"United States\"},\n {\"order_id\": 1176, \"total\": 44.0, \"channel\": \"web\", \"country\": \"United States\"},\n]\n# 1. Revenue per country, with country names cleaned by .strip().title()\nrevenue_by_country = {}\n\n# 2. order_id of every order with a total of 60 or more, in the original order\nlarge_ids = []\n\n# 3. How many orders came through the \"web\" channel\nweb_count = 0\n","solution":"orders = [\n {\"order_id\": 2574, \"total\": 65.0, \"channel\": \"mobile app\", \"country\": \"Canada\"},\n {\"order_id\": 2069, \"total\": 18.0, \"channel\": \"mobile app\", \"country\": \"United States\"},\n {\"order_id\": 1200, \"total\": 33.0, \"channel\": \"mobile app\", \"country\": \"united kingdom\"},\n {\"order_id\": 1809, \"total\": 35.0, \"channel\": \"web\", \"country\": \"United States\"},\n {\"order_id\": 1434, \"total\": 72.0, \"channel\": None, \"country\": \"Saudi Arabia\"},\n {\"order_id\": 2089, \"total\": 58.0, \"channel\": \"marketplace\", \"country\": \"UNITED KINGDOM\"},\n {\"order_id\": 2490, \"total\": 89.0, \"channel\": \"web\", \"country\": \"United Arab Emirates\"},\n {\"order_id\": 2985, \"total\": 24.0, \"channel\": \"web\", \"country\": \"United Kingdom\"},\n {\"order_id\": 2544, \"total\": 58.0, \"channel\": \"mobile app\", \"country\": \"Jordan\"},\n {\"order_id\": 2375, \"total\": 329.0, \"channel\": \"mobile app\", \"country\": \"United Kingdom\"},\n {\"order_id\": 1129, \"total\": 29.0, \"channel\": \"web\", \"country\": \"United States\"},\n {\"order_id\": 1176, \"total\": 44.0, \"channel\": \"web\", \"country\": \"United States\"},\n]\nrevenue_by_country = {}\nfor order in orders:\n country = order[\"country\"].strip().title()\n revenue_by_country[country] = revenue_by_country.get(country, 0) + order[\"total\"]\nlarge_ids = [o[\"order_id\"] for o in orders if o[\"total\"] >= 60]\nweb_count = 0\nfor o in orders:\n if o[\"channel\"] == \"web\":\n web_count += 1\n","checks":[{"id":"countries-clean","test":"sorted(revenue_by_country) == ['Canada', 'Jordan', 'Saudi Arabia', 'United Arab Emirates', 'United Kingdom', 'United States']"},{"id":"uk-total","test":"round(revenue_by_country.get('United Kingdom', 0), 2) == 444.0"},{"id":"all-revenue","test":"round(sum(revenue_by_country.values()), 2) == 854.0"},{"id":"large-ids","test":"list(large_ids) == [2574, 1434, 2490, 2375]"},{"id":"web-count","test":"web_count == 5"}],"title":"Summarise twelve orders by country","prompt":"orders holds twelve rows from the messy export, and the country names have inconsistent capitals. Build revenue_by_country, a dictionary from cleaned country name (use .strip().title()) to the sum of total. Build large_ids, a list of the order_ids whose total is 60 or more, in their original order. Set web_count to the number of orders whose channel is exactly "web".
orders once for the dictionary: clean the country into a variable first, then update the total for that key.","revenue_by_country[country] = revenue_by_country.get(country, 0) + order["total"] adds to an existing total or starts a new one.","large_ids = [o["order_id"] for o in orders if o["total"] >= 60]. For web_count, sum(1 for o in orders if o["channel"] == "web") or a loop with += 1 both work."],"explanation":"Cleaning the key before using it is what makes "united kingdom", "UNITED KINGDOM" and "United Kingdom" land in the same bucket; without it you would get three UK entries and a total split three ways. The .get(key, 0) pattern handles the first appearance of each country. The comprehension keeps the list in the original order because it walks orders front to back. The grand total across countries is still 854.0, which is a quick way to confirm that grouping neither lost nor double-counted a row: always check that the parts add up to the whole.
large_ids is [2574, 1434, 2490, 2375]","web-count":"web_count is 5"}},"words":716,"readMinutes":3}