3. Operators & control flow

Beginner · 8 min read

Control flow decides which code runs and how many times.

3.1 Comparison and logical operators

score = 0.82
print(score > 0.8)                  # → True
print(score == 0.82, score != 1)    # → True True
print(0.5 <= score <= 1.0)          # → True   comparisons can be chained

has_key, is_admin = True, False
print(has_key and not is_admin)     # → True   and / or / not
print("rag" in "retrieval-augmented rag")  # → True   membership

3.2 if / elif / else

tokens = 4200

# Python uses indentation (4 spaces) to mark the block that belongs to each branch.
if tokens > 8000:
    print("Too long — split it")
elif tokens > 4000:
    print("Long — consider a bigger model")   # → Long — consider a bigger model
else:
    print("Fits easily")

# One-line form (conditional expression)
size = "large" if tokens > 4000 else "small"
print(size)                                     # → large

3.3 for loops

models = ["gpt-4o-mini", "llama-3.1-8b", "claude"]

for model in models:                 # loop over any sequence
    print(model)

for i in range(3):                   # range(3) → 0, 1, 2
    print("attempt", i)

for i, model in enumerate(models, start=1):   # index + value together
    print(i, model)                  # first line → 1 gpt-4o-mini

3.4 while loops, break and continue

attempt = 0
while attempt < 5:                   # repeat while the condition is True
    attempt += 1
    if attempt == 2:
        continue                     # skip the rest of this round
    if attempt == 4:
        break                        # leave the loop completely
    print("try", attempt)            # prints try 1, then try 3

3.5 match (Python 3.10+)

def describe(status):
    match status:                    # compare one value against several patterns
        case 200:
            return "ok"
        case 429:
            return "rate limited — retry later"
        case 500 | 502 | 503:
            return "server error"
        case _:                      # _ matches anything else
            return "unknown"

print(describe(429))                 # → rate limited — retry later

Why it matters for GenAI

Retrying an LLM call on a rate limit (HTTP 429) is a while loop with a break on success — you'll write this pattern constantly.

Practice

• Loop over the numbers 1–20 and print "chunk" for multiples of 5, otherwise the number.
• Write a while loop that "retries" up to 3 times and stops early when attempt == 2.

Answer

for n in range(1, 21):
    print("chunk" if n % 5 == 0 else n)