Filtering: count only the cards you want

A filter is a condition that selects the records relevant to a question. A filtered iteration still visits the records systematically, but only matching records update the chosen answer.

Two ways to filter #

You can first separate matching cards into a pile and then count them. Or you can check each card and count a match immediately, in one scan. Both methods give the same selected set if the condition is unchanged.

VerbCount ← 0
For each word card X:
    If X.category = "Verb":
        VerbCount ← VerbCount + 1
    Mark X as visited
Report VerbCount

Trace the teaching words #

The answer is 2 verbs among 7 visited word cards. The lecture’s own paragraph has reported counts of 7 verbs and 7 pronouns; do not confuse its totals with this small teaching dataset.

Several counters can share one scan #

Initialize VerbCount and PronounCount to zero. Update the appropriate count from each card’s recorded category. If you also count Other, the three totals should add to the number of visited cards, assuming each card has exactly one category label.

Tutorial example: all food versus no food #

Count bills, not individual food items. For a nonempty bill, “all food” means every item is food. “No food” means none of its items are food. A mixed bill meets neither condition. “Not all food” includes both mixed and no-food bills, so it is not the same as “no food”.

Assume the tutorial bills contain at least one item. If empty bills are possible in another task, explicitly define their handling before using all/none conditions; excluding or flagging an empty bill avoids ambiguity in this exercise.

Source: Lecture 1.2; Tutorial 1.1. Small teaching datasets and diagrams are prepared for these notes; lecture and tutorial totals are identified explicitly.