TSV SQL Viewer

一个 VS Code 扩展原型，用 DuckDB SQL 查询和更新 TSV 文件。
An experimental VS Code extension for querying and updating TSV files with DuckDB SQL.

中文

当前能力

• *.tsv 默认仍按普通文本打开
• 可通过编辑器右上角入口切换到 TSV Query 视图
• 内置 DuckDB（@duckdb/node-api）执行 SELECT / UPDATE / DELETE / INSERT
• 更新先作用于当前会话，点击“写回 TSV”后覆盖原文件
• 自动把原始表头映射成安全的 SQL 列名，并在界面里展示映射关系
• 查询结果只预读前 200 行，避免一次性把大结果集全量拉进界面
• 支持基础列类型推断，目前会识别 BOOLEAN / BIGINT / DOUBLE / TEXT

开发

npm install
npm run build

然后在 VS Code 中运行扩展开发宿主：

1. 打开本项目
2. 执行 Run Extension
3. 在新窗口中打开任意 *.tsv 文件
4. 点击编辑器右上角的 Open TSV Query View

使用方式

表名固定为：

tsv_data

示例查询：

SELECT *
FROM tsv_data
ORDER BY name DESC
LIMIT 200;

示例更新：

UPDATE tsv_data
SET status = 'done'
WHERE owner = 'sam';

执行更新后，点击界面中的 Write Back TSV 即可持久化。

当前限制

• 当前版本仍基于 CustomTextEditor，VS Code 打开文件时依然会先把整个 TSV 作为文本载入
• 写回 TSV 时仍需要把当前结果完整导出一次
• 查询结果展示最多渲染前 200 行
• 列类型推断是保守策略，遇到混合格式会自动降级成 TEXT
• 下一步应该把文档模型改成真正面向大文件的自定义文档，而不是直接依赖文本编辑器

English

Features

• *.tsv files still open as normal text files by default
• You can switch to the TSV Query view from the editor title action
• Uses DuckDB (@duckdb/node-api) to run SELECT / UPDATE / DELETE / INSERT
• Updates are first applied in the current session, then written back to the original TSV file
• Automatically maps raw TSV headers to safe SQL column names and shows the mapping in the UI
• Query results are previewed with a 200-row cap to avoid rendering very large result sets at once
• Includes basic column type inference for BOOLEAN / BIGINT / DOUBLE / TEXT

Development

npm install
npm run build

Then run the extension development host in VS Code:

1. Open this project in VS Code
2. Start Run Extension
3. Open any *.tsv file in the new window
4. Click Open TSV Query View in the editor title area

Usage

The table name is fixed as:

tsv_data

Example query:

SELECT *
FROM tsv_data
ORDER BY name DESC
LIMIT 200;

Example update:

UPDATE tsv_data
SET status = 'done'
WHERE owner = 'sam';

After running an update, click Write Back TSV in the query view to persist the changes back to the TSV file.

Current Limitations

• The current version still relies on CustomTextEditor, so VS Code reads the whole TSV file as text first
• Writing changes back still requires exporting the current in-memory result to a full TSV file
• The result preview is capped at 200 rendered rows
• Type inference is conservative and falls back to TEXT when a column contains mixed formats
• The next major step should be moving to a true large-file document model instead of relying on the text editor pipeline