连接池与事务
使用连接池高效管理数据库连接，并在 next.jdbc 中通过事务确保数据完整性。
连接池与事务 是 CoddyKit 上的免费 Clojure Functional Programming & JVM Backend Development 课时。 这是第 4 节课，共 4 节。 你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。 这是 Clojure Functional Programming & JVM Backend Development 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 Clojure Functional Programming & JVM Backend Development 课程共包含 4 节课。
本课时的部分内容尚未翻译，以英文显示。
Why Connection Pooling?
Opening a new database connection for every query is slow and resource-heavy. A connection pool keeps a set of reusable connections ready, dramatically improving throughput.
The HikariCP Pool
The most popular JVM connection pool is HikariCP. The next.jdbc.connection namespace provides a helper to build a pooled datasource.
(require '[next.jdbc.connection :as connection])
(import '[com.zaxxer.hikari HikariDataSource])Creating a Pooled Datasource
Use connection/->pool with the pool class and your db spec. The returned datasource is what you query against.
(def datasource
(connection/->pool HikariDataSource
{:dbtype "postgresql"
:dbname "mydb"
:username "app"
:password "secret"}))Querying Through the Pool
Once created, pass the pooled datasource directly to jdbc/execute! just like a plain connection. The pool checks a connection out and back automatically.
(require '[next.jdbc :as jdbc])
(jdbc/execute! datasource
["SELECT * FROM users WHERE active = ?" true])Closing the Pool
A pooled datasource is a closeable resource. Close it on shutdown to release all connections.
(.close datasource)What Is a Transaction?
A transaction groups several statements so they all succeed or all fail together. This guarantees consistency, like transferring money between accounts.
with-transaction
next.jdbc provides with-transaction. Statements inside the body run atomically; if an exception is thrown, everything rolls back.
(jdbc/with-transaction [tx datasource]
(jdbc/execute! tx ["UPDATE acct SET bal = bal - 100 WHERE id = 1"])
(jdbc/execute! tx ["UPDATE acct SET bal = bal + 100 WHERE id = 2"]))Automatic Rollback
If any statement throws, the transaction is rolled back automatically and the exception propagates. No partial updates are committed.
(jdbc/with-transaction [tx datasource]
(jdbc/execute! tx ["INSERT INTO orders (id) VALUES (1)"])
(throw (ex-info "boom" {}))) ; insert above is rolled backIsolation Levels
You can specify an isolation level to control how concurrent transactions see each other.
:read-committed:repeatable-read:serializable
(jdbc/with-transaction [tx datasource {:isolation :serializable}]
(jdbc/execute! tx ["UPDATE counters SET n = n + 1 WHERE id = 1"]))Manual Rollback
You can force a rollback without throwing by setting the rollback flag on the connection.
(jdbc/with-transaction [tx datasource]
(jdbc/execute! tx ["INSERT INTO logs (msg) VALUES ('test')"])
(.rollback tx))Pooling + Transactions Together
In production you create one pooled datasource at startup and wrap critical multi-step writes in with-transaction. This gives both performance and integrity.
(defn transfer! [ds from to amount]
(jdbc/with-transaction [tx ds]
(jdbc/execute! tx ["UPDATE acct SET bal = bal - ? WHERE id = ?" amount from])
(jdbc/execute! tx ["UPDATE acct SET bal = bal + ? WHERE id = ?" amount to])))Quick Check
Test your transaction knowledge.
Recap
You learned to use connection pooling and transactions with next.jdbc.
- Build a pool with
connection/->pooland HikariCP - Wrap atomic writes in
with-transaction - Failures roll back automatically; choose isolation levels for concurrency
常见问题解答
「连接池与事务」课时是免费的吗？
是的 — 「连接池与事务」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 Clojure Functional Programming & JVM Backend Development 课程的其余内容，请升级到 CoddyKit PRO。 Clojure Functional Programming & JVM Backend Development 课程共包含 4 节课。
「连接池与事务」这节课中我会学到什么？
使用连接池高效管理数据库连接，并在 next.jdbc 中通过事务确保数据完整性。 你通过在浏览器中直接运行的动手代码来练习 Clojure Functional Programming & JVM Backend Development，全天候 AI 导师会在你学习这节课的过程中回答你的问题。
学习 Clojure Functional Programming & JVM Backend Development 需要有经验吗？
无需任何先前经验。CoddyKit 上的 Clojure Functional Programming & JVM Backend Development 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。 这是第 4 节课，共 4 节。
「连接池与事务」课时需要多长时间？
大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。
我能在这节 Clojure Functional Programming & JVM Backend Development 课中编写并运行代码吗？
能。每节 Clojure Functional Programming & JVM Backend Development 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。