Ghost in the SQL Data Demo Walkthrough: Solve Your First Case Step by Step
A complete Ghost in the SQL Data demo walkthrough covering the core investigation loop, SQL query patterns, common errors, and Expert mode tips.
If you have ever wanted to solve a crime with a keyboard instead of a gun, this Ghost in the SQL Data demo walkthrough is the place to start. The free Steam trial drops you into Paris as a detective whose only weapon is SQL, and it expects you to filter millions of emergency-call records down to a defensible conclusion. That sounds intimidating, but the demo is built to teach you, and this Ghost in the SQL Data demo walkthrough breaks the whole loop into repeatable steps you can apply to every case.
What the Ghost in the SQL Data Demo Actually Covers
The demo is a standalone free trial of Ghost in the SQL Data, a detective game where investigations run on real queries against emergency-call and city databases. There is no action sequence and no chase scene. Every case asks you to interpret an objective, locate the relevant records, and connect evidence into an answer the game will accept.
The full release launched on Steam on October 29, 2026, with five cases and multiple SQL learning checkpoints. The demo exists to teach the fundamental loop before you commit to the paid product: translate a case prompt into database questions, filter a huge dataset down to a handful of meaningful rows, and verify that your result actually answers what was asked.
| Demo Aspect | Details |
|---|---|
| Platform | Steam (free standalone demo) |
| Core mechanic | SQL queries against Paris emergency-call data |
| Modes | Guided and Expert |
| Full release | October 29, 2026, 5 cases |
| Community reception | Roughly 94% positive across 94 Steam reviews |
Community reports suggest the demo is short but dense. You are not meant to speedrun it. You are meant to leave with working habits that carry into the full caseload.
Start in Guided Mode, Then Graduate to Expert
The single most common piece of advice from player experience is simple: do not start in Expert mode to prove something. Guided mode walks you through the investigation workflow step by step, covering SELECT, WHERE, and basic filtering. It tells you what kind of question to ask before you ask it.
Expert mode removes that scaffolding. It gives you less structure, which is exactly why it is the better second playthrough and the worse first one. If you already write SQL professionally, jumping straight to Expert is reasonable, but even then the Guided introduction case is worth finishing because it teaches the game's specific interface conventions.
| Mode | Structure Level | Best For |
|---|---|---|
| Guided | High — step-by-step prompts | First-time players, SQL beginners |
| Expert | Low — minimal hints | Players who know SELECT, WHERE, JOIN |
A useful middle path: play the introduction case in Guided mode, then replay it in Expert mode before touching anything else. You will notice how much of the reasoning you were outsourcing to the prompts.
The Four Investigation Patterns You Will Meet
Every case in the demo and the full game is built from a small set of repeatable patterns. Recognizing which pattern a mission uses tells you which SQL clauses you need before you type a single character. This is the core insight of any Ghost in the SQL Data demo walkthrough worth reading.
- Identity Search — isolate the correct person among many records using name, status, and ID fields. Beginner-friendly, teaches basic filtering with SELECT and WHERE.
- Timeline Reconstruction — rebuild what happened and when using date filters and ORDER BY. Intermediate difficulty.
- Cross-Table Connection — link separate records through shared identifiers using JOIN on person_id or location_id. Intermediate to advanced.
- Evidence Filtering — combine multiple conditions to find useful records using WHERE, AND, and date ranges. Advanced compound queries.
| Pattern | Key SQL Clauses | Difficulty | Demo Coverage |
|---|---|---|---|
| Identity Search | SELECT, WHERE | Beginner | Core teaching example |
| Timeline | WHERE, ORDER BY | Intermediate | Date filter practice |
| Cross-Table | JOIN, shared keys | Intermediate+ | Relational reasoning |
| Evidence Filtering | WHERE, AND, ranges | Advanced | Compound queries |
The most common reason players get stuck in the demo is misclassifying the pattern. Trying to solve a timeline case with identity-only filters will leave you with a correct suspect and no way to prove anything. Classify first, then query.
A Five-Stage Walkthrough for Every Demo Case
Work through each case with the same rhythm. Each stage narrows the gap between the case prompt and a defensible answer, and the repetition is what makes the later cases feel easy.
Stage 1 — Read the case objective. Separate requested facts from story flavor. Underline names, dates, places, and relationships, then turn the objective into one or more concrete database questions. If the prompt asks "who called from the scene," your question is "which rows match this location and this time window."
Stage 2 — Inspect the available data. Identify which table holds the information you need. Check column names and run a small, limited query first to understand what each field represents. Guessing at column names is the fastest way to waste ten minutes.
Stage 3 — Build a focused query. Use SELECT to choose visible evidence and WHERE to narrow records. Add conditions one at a time with AND so you can watch how each clue changes the candidate set.
Stage 4 — Compare and interpret results. Treat results as evidence, not as automatic answers. Check repeated values, meaningful differences, and whether every required condition is satisfied before you conclude anything.
Stage 5 — Connect and verify. If the answer needs data from another source, JOIN through the shared identifier. Re-read the objective and confirm the final result addresses every part of the question.
| Stage | Main Action | Typical Clauses |
|---|---|---|
| 1. Read objective | Convert prompt into database questions | None yet |
| 2. Inspect data | Identify tables and columns | SELECT with a limit |
| 3. Build query | Narrow records incrementally | SELECT, WHERE, AND |
| 4. Interpret | Compare and validate evidence | ORDER BY, aggregates |
| 5. Verify | Join sources, re-check objective | JOIN, shared keys |
Incremental queries win. Adding one filter at a time keeps every result explainable. If a query fails or returns nothing, you know exactly which condition broke it instead of debugging five changes at once.
Common Query Errors and How to Fix Them
Most demo failures come from a handful of SQL mistakes rather than wrong detective theory. Match your symptom to the fix below before you reconsider your investigation.
| Error Symptom | Likely Cause | Fix |
|---|---|---|
| Query fails near a name | Text value not quoted | Use single quotes: name = 'Alice' |
| Date query returns zero rows | Wrong date format | Use ISO YYYY-MM-DD format |
| Filter returns nothing | Mismatched operator | Use a range for periods, not = on one day |
| Result set far too large | AND vs OR confusion | Use IN (...) for one-of-several values |
The rule of thumb: check structure before theory. When a query errors or returns empty results, verify table names, field names, quotes, and date formats first. Most of the time your investigation theory was correct and only the query structure was wrong. If you want to sharpen your SQL fundamentals outside the game, the W3Schools SQL tutorial is a solid free reference for the clauses this game uses.
Demo Progression Checklist and Expert Mode Tips
Before you move from the demo to the full release, complete these milestones. Each one proves a specific investigation skill that later cases assume you already have.
- Complete the Guided mode introduction case
- Write a working SELECT and WHERE query unaided
- Rebuild an event timeline using date filters and ORDER BY
- Connect two tables with a JOIN on a shared identifier
- Finish one case in Expert mode without hints
In Expert mode, keep a written clue log. List confirmed clues and ruled-out suspects as you go. The game gives less structure there, and a manual log prevents you from repeating queries you have already disproven — a habit that pays off across all five cases in the full release.
| Milestone | Skill Proven |
|---|---|
| Guided intro complete | Understands the loop |
| Unaided SELECT/WHERE | Basic filtering |
| Timeline rebuild | Date logic and ordering |
| Two-table JOIN | Relational reasoning |
| Expert case finished | Independent investigation |
One more Expert mode tip from community reports: when a case stalls, re-read the objective before rewriting your query. Players tend to assume the data is wrong when the actual problem is that they answered a slightly different question than the one asked.
Demo vs Full Release: What You Get
The demo is designed primarily as an introduction to the investigation format. Treat it as practice for the query workflow rather than a prologue whose save data transfers to the full release.
| Feature | Demo | Full Release |
|---|---|---|
| Access | Free on Steam | Paid Steam product |
| Content | Introductory investigation loop | 5 cases plus SQL learning checkpoints |
| Modes | Guided and Expert available | Full progression experience |
| Best for | Testing the concept | Players who want the complete caseload |
If you finish the demo checklist and still want more casework, the full game is the natural next step. If you bounce off SQL entirely, the demo has done its job by telling you early.
FAQ
Is the Ghost in the SQL Data demo free?
Yes. The demo is a free standalone Steam product that introduces the core SQL-driven investigation loop. The full release is a paid product with five cases and additional SQL learning checkpoints.
Do I need to know SQL before playing the demo?
No. Guided mode introduces the investigation workflow step by step, covering SELECT, WHERE, and basic filtering. Players with SQL experience can jump into Expert mode for a more independent investigation.
Does demo progress carry over to the full game?
The demo is built as an introduction to the investigation format. Treat it as practice for the query workflow rather than a prologue whose save data transfers to the full release.
What should I do if my query returns no results?
Check the query structure first: single quotes around text values, ISO YYYY-MM-DD date formats, correct comparison operators, and shared key fields in JOINs. Only revisit your investigation theory after confirming the structure is correct. Following this Ghost in the SQL Data demo walkthrough in order will prevent most of these errors before they happen.