Decisions and loops: if, for and while

Open a real orders file and you will not find 500 identical rows.

Some orders are paid, some are cancelled, and a few contain missing or negative amounts.

When you download a daily export from a store, rows arrive in mixed states.

If your pipeline treats every row the same way, it either crashes or quietly corrupts financial reports.

Conditionals let code choose what to do with a record, while loops repeat the decision across every row.

An if statement inspects one row, while a for loop applies that rule to thousands of rows automatically.

Write the decision rule once. Let Python repeat it across every row.

Data pipelines require specific helpers to number rows, unpack dictionaries, and handle retries cleanly.

Rather than writing manual counter variables, Python standard helpers give you clean tools for each task.

Each helper eliminates manual counter variables and keeps pipeline logic concise and readable.

Mastering these helpers allows you to write robust parsers that handle edge cases gracefully.

Pipelines process millions of records that arrive on varying schedules.

Leadership asks for yesterday paid revenue and cancelled count from 2 lakh orders.

Writing 2 lakh hardcoded lines breaks the moment tomorrow file arrives with a different row count.

A compact loop applies the exact same business logic whether a file has 10 rows or 10 million rows.

When tomorrow batch arrives with extra records, the loop processes them without code modifications.

In addition, iterating record by record consumes minimal computer memory compared to loading uncompressed spreadsheets.

A loop decouples pipeline logic from the number of incoming rows.

Python gives you clear control structures to evaluate conditions and iterate collections.

An if statement tests a condition, executing indented lines only when the condition evaluates to True.

Add elif for secondary checks and else for remaining cases.

Predict which branch prints when status is paid before checking the output below.

The first condition was true, so Python executed that branch and skipped all remaining checks.

Using elif ensures only one branch runs, whereas separate if statements would test every single check independently.

Two equals signs ask whether two values match. A single equals sign assigns a value.

Using a single equals sign in an if condition triggers an immediate syntax error.

Python warns you that an assignment was attempted where an equality comparison was expected.

Joining conditions with and and or allows you to build sophisticated data validation filters.

Remember that comparisons evaluate from left to right, and parentheses clarify complex logic.

Python requires indented lines under every colon to know where branches start and stop.

Forgetting the four-space indentation halts execution immediately with an error.

Notice that Python points directly to the line where an indented block was expected.

Every branch line ends with a colon and must be indented by four spaces.

Most pipeline datasets arrive as lists of dictionaries, where each dictionary represents one record.

The for loop pulls one record at a time, making it accessible through a named variable.

On each turn, the order variable binds to the next dictionary until all records have been processed.

You extract specific fields using dictionary keys, such as order['total'].

In data engineering, you frequently need both row index and record content to report error line numbers.

The enumerate helper supplies both, and start=1 sets human-readable counting.

Predict what prints if you omit start=1: does the first row number 0 or 1?

Without start=1, Python starts numbering from index 0.

When a malformed record appears on row 482, enumerate lets your log statement pinpoint the exact line.

Iterating a dictionary directly yields only keys. Adding .items() yields key and value together.

This is standard practice when reading environment variables or summary metric tables.

Two variables unpack the pair on every turn, perfect for reading configuration files and metric summaries.

If you only need values, Python provides .values(), but .items() delivers the complete picture.

When you need a loop to execute N times without an existing list, use range.

This is common when splitting large datasets into fixed batch chunks.

Predict the printed numbers before looking below: does it start at 0 or 1?

range(3) produces three numbers starting at 0 and stopping before 3.

To produce numbers from 1 to 3, pass two arguments: range(1, 4).

A for loop iterates known collections, while a while loop repeats as long as a condition remains True.

Always include an internal counter advancement so retry loops terminate safely.

When polling external APIs, bounded while loops prevent stalled jobs from hanging indefinitely.

Forgetting to advance the counter leaves the condition True forever, creating a runaway infinite loop.

Every while loop must advance its own termination condition.

Use continue to skip remaining steps for a bad row and break to exit the loop upon finding a match.

Here is a batch where one order has a missing total. We want the id of the first order exceeding 500.

Predict which order id prints, and whether order 104 is ever evaluated.

Order 102 was skipped by continue, order 103 triggered break, and order 104 was never evaluated.

Without the continue check, comparing None with a number triggers an immediate crash:

Checking for None before comparisons protects pipelines against malformed inputs.

In production, continue filters out bad rows while break short-circuits search operations.

Now combine conditionals and loops into standard data engineering patterns.

Initialize totals at zero before the loop, then route rows into their respective accumulators.

Starting totals at 0.0 establishes floating-point arithmetic from the start.

Predict the final paid revenue and cancelled count before checking the output below.

Priya order was cancelled, so her 300 amount never touched paid_total; only the cancelled count incremented.

Tracing variable states turn by turn reveals exactly how conditions partition metrics.

When an order contains a negative or missing value, separate clean rows from bad rows using two lists.

This two-list approach prevents pipeline failure while preserving corrupted records for engineering review.

Predict how many rows land in bad, and which customer order is quarantined.

Order 102 had a negative price, so it moved into quarantine whole without halting pipeline execution.

Never call orders.remove(order) during a loop; deleting elements while iterating shifts indices and skips records.

Building separate clean and quarantine lists operates in linear time without index corruption bugs.

Four rules to remember for loops in data engineering:

1. Use for loops to iterate across collections and while loops for bounded polling retries.
2. Always check for None before comparing values to prevent runtime TypeErrors.
3. Accumulate metrics into initialized variables before the loop rather than recomputing them.
4. Split malformed records into a quarantine list rather than modifying the iterated list in place.

In the next lesson you will learn how functions encapsulate looping logic into reusable, testable components.

You will also learn how try and except catch runtime errors so unexpected values never crash production pipelines.

In the exercise you get a list of customer transactions. Write a for loop that checks each status with if and elif.

Add up the revenue from paid orders, count cancelled orders, count unknown statuses, and store all three in a dictionary named result.