title	Anki Decks
type	index
tags	anki spaced-repetition flashcards
status	published

Anki Decks

Spaced-repetition flashcards distilled from the cheat sheets, fundamentals, and key-takeaway sections of this guide. Use these the last 2–4 weeks before your exam to lock in high-frequency facts.

How the system works

The repo stores deck source as markdown (human-readable, Obsidian-friendly, version-controlled). A small Python script converts each source file into an Anki-importable TSV under anki/build/ which is .gitignored. The TSV is what you actually import into the Anki desktop app.

anki/
├── README.md                          # this file
├── format.md                          # deck source format spec
├── build.py                           # Python stdlib-only converter
├── decks/                             # source decks (committed)
│   └── shared/
│       ├── delta-lake.md              # Delta Lake — 27 cards
│       └── unity-catalog.md           # Unity Catalog — 22 cards
└── build/                             # generated output (gitignored)
    └── shared/
        ├── delta-lake.txt
        └── unity-catalog.txt

Available decks

Deck	Source	Cards	Best for
Delta Lake	decks/shared/delta-lake.md	27	DE Associate, DE Professional
Unity Catalog	decks/shared/unity-catalog.md	22	every cert

More decks coming as the community contributes — see Contributing a new deck below.

Workflow — for learners

1. Install Anki (free, desktop, all platforms)

2. Clone this repo locally

3. Build the TSV files:

   python3 anki/build.py

   No pip install needed — Python 3.9+ standard library only.

4. Import into Anki: open Anki → File → Import → select each .txt file under anki/build/. Anki reads the #deck: line at the top of each file, so each deck lands in the right hierarchy (Databricks::DE Associate::Delta Lake, etc.).

5. Study daily. Anki's default settings (new: 20/day, review: unlimited) work fine for cert prep.

Tip

If you change a deck source file and re-run python3 anki/build.py, Anki's import dialog will offer to update existing notes rather than duplicate them. Pick that option.

Workflow — for the curious (no Anki)

If you don't want to install Anki, the source markdown files are usable as quiz material directly:

• In Obsidian: the > [!success]- Answer callout is foldable — read the question, think, then click to reveal the answer.
• In GitHub: the callout renders as an open block, so just scroll past the answer to skip ahead.

The Anki workflow is just the higher-leverage option because spaced repetition genuinely beats one-pass reading for memorising lookups like "default VACUUM retention" or "Z-order vs Liquid Clustering."

Contributing a new deck

PRs welcome. Process:

1. Pick a topic that maps to an existing cheat sheet (e.g., shared/cheat-sheets/mlflow-quick-ref.md → anki/decks/shared/mlflow.md) — concentrated reference content is what makes the best flashcards.
2. Follow the format spec. Each card has a clear front (one question or scenario) and a precise back (the answer, with citation/qualifier where appropriate).
3. Aim for 20–35 cards per deck — large enough to be useful, small enough to ship a quality deck per PR.
4. Run python3 anki/build.py --check to validate the format.
5. Run python3 anki/build.py to generate output and spot-check a few cards in Anki before opening the PR.
6. Add a row to the "Available decks" table above.
7. Submit your PR.

Card-writing principles

• One fact per card. "What is X?" beats "What are the five properties of X?" — split the second into five cards.
• Front is the question, not the topic. "Define ACID" → bad. "Which Delta Lake guarantee ensures all-or-nothing transaction visibility?" → good.
• Back is just the answer, not a re-statement of the question. Anki shows the front above the back; redundancy wastes brain time.
• Be specific in the back. "Use OPTIMIZE" → not enough. "OPTIMIZE <table> — coalesces small files into the table's delta.targetFileSize (default 1 GB)." → enough.
• Code in code spans. Use backticks for SQL, Python, file paths, and config keys. The builder preserves them.
• Avoid trivia. Skip cards about exact API parameter names that change every release. Focus on conceptual + decisional questions ("when to use X vs Y", "what does X guarantee", "what's the default for Y").
• Mark deprecated content explicitly. If a card references something current as of the May 2026 / Nov 2025 / Mar 2026 blueprint, name the cutoff in the answer.

What this isn't

• Not a replacement for the topic files — flashcards lock in facts, not understanding. Read the topic file first.
• Not a substitute for mock exams — flashcards train recall, not exam pacing or distractor-elimination. Use both.
• Not a full deck per cert (yet). Contributions wanted.

License

Same as the rest of the repo: MIT. Cards are released under the same terms; you're free to import them into your own Anki and use them however.

---

← Back to repo root | Deck format spec →