# Read the Manual > Some titles promise to walk you through it. Count the ones that say so out loud. Canonical URL: Domain: SQL · Difficulty: easy · Seniority: L3 ## Problem The content team wants to know how much of the catalog is explicitly framed as a "how-to" walkthrough. Count how many content items have the word 'how' somewhere in their title (case-insensitive). ## Worked solution and explanation ### It looks trivial. Two ways to quietly get it wrong. Strip the costume and this is a case-insensitive substring count, and the whole game is what 'contains how' actually means. There are two quiet ways to blow it. Match the standalone word (or worse, `title = 'how'`) and you drop every 'How to Data Science' the content team is actually asking about. Match the substring but forget to fold case, and a bare `LIKE '%how%'` silently skips every capitalized 'How', which is most of them. The count only lands when you treat 'how' as any run of h-o-w, anywhere in the title, regardless of case. Nobody sweats this on a whiteboard; the tell of a junior answer is one that assumes the titles are clean and single-cased. --- ### Break down the requirements #### Step 1: Match titles containing 'how' `WHERE LOWER(title) LIKE '%how%'` finds all content items with 'how' anywhere in the title, regardless of case. The leading and trailing `%` are what make it a substring test rather than an exact match, and LOWER() is what keeps 'How' from slipping through. #### Step 2: Count the matches `COUNT(*)` with no GROUP BY collapses the matches to a single number, which is exactly the one-row, one-column shape the content team wants. --- ### The solution **Case-insensitive pattern match with count** ```sql SELECT COUNT(*) AS howto_count FROM content_items WHERE LOWER(title) LIKE '%how%' ``` > **Cost Analysis** > > Full scan of 3M rows with a LIKE pattern. The leading `%` prevents index usage, and wrapping `title` in LOWER() also blocks any plain index on the column. For frequent queries, a full-text index or a computed column storing LOWER(title) would help. > **Interviewers Watch For** > > Whether the candidate handles case at all, and whether they do it portably. SQLite's LIKE is case-insensitive for ASCII by default, so a bare `LIKE '%how%'` passes here and hides the bug; the same query undercounts on PostgreSQL, where LIKE is case-sensitive. Wrapping the column in LOWER() is the answer that survives the move to production. > **Common Pitfall** > > Using `= 'how'` instead of `LIKE '%how%'`. The exact match would miss titles like 'How to Write SQL Joins' or 'A guide on how to optimize queries'. --- ## Common follow-up questions - What if you needed the count per content_type? _(Add GROUP BY content_type to break down how-to titles by format.)_ - How would you handle titles with HTML entities like '&how'? _(Tests data cleaning awareness before pattern matching.)_ - What if the search needed to match only the whole word 'how' and not substrings like 'showcase'? _(The LIKE '%how%' pattern matches 'how' as a substring, so it also catches words like 'showcase' or 'somehow' that contain the letters; matching the standalone word would need word-boundary handling.)_ ## Related - [All practice problems](https://datadriven.io/problems) - [Mock interview mode](https://datadriven.io/interview/read_the_manual) - [SQL Interview Questions](https://datadriven.io/sql-interview-questions) - [Data Engineering Interview Prep Guide](https://datadriven.io/data-engineer-interview-prep) - [Daily Challenge](https://datadriven.io/daily) --- Source: DataDriven (https://datadriven.io). DataDriven is the data engineering interview community. Live code execution in SQL, Python, and Spark sandboxes. Every feature is open to every member.