Technology

Elasticsearch Toolbox

storagesearchtier2
On this page

Interview Trigger

看到这些需求，第一反应是 Elasticsearch：

Search / Keyword Search / Full-text Search
Autocomplete / Fuzzy Search / Ranking / Filter + Search

What

A distributed search engine built on an inverted index.

关键词：Distributed / Full-text Search / Inverted Index

为什么快？

1. Inverted Index

最重要的原因。不是 document → terms，而是 term → documents。

2. Tokenization

"Bitcoin is amazing" → [Bitcoin, is, amazing] → 建立 Index

3. Distributed

Index 可以 Sharding：Shard1 / Shard2 / Shard3，多个 Node 一起 Search。

支持的搜索类型

  • Full-text Search: bitcoin wallet
  • Prefix Search: 输入 bit → 找到 Bitcoin
  • Fuzzy Search: 输入 bitcon（少一个 i）→ 仍然找到 Bitcoin
  • Ranking: 按 Relevance 排序，不是 Primary Key

高频 Follow-up

Q1: Why Elasticsearch instead of MySQL?

Elasticsearch is optimized for full-text search using inverted indexes, while MySQL indexes are designed primarily for exact lookups and range queries.

Q2: What is an inverted index?

Instead of mapping a document ID to its content, an inverted index maps each term to the list of documents containing that term, allowing very fast keyword searches.

Q3: Should Elasticsearch be the source of truth?

No. I’d use MySQL as the source of truth and keep Elasticsearch synchronized asynchronously through CDC or Kafka.

Q4: When would you NOT use Elasticsearch?

Transfer / Wallet / Payment — 这些需要 Transaction 和 Strong Consistency，还是 MySQL。