Bamboo 0.2.2

Bamboo 0.2.2 adds query history and named saved queries.

• Open the searchable Query history dialog from the right edge of the editor. History records successful manual runs; Saved keeps queries you name for reuse, with rename and delete actions.
• Open a query in its original source's editor without running it. Draft replacement requires confirmation, and queries from another source cannot replace the current editor.
• History and saved queries persist privately on this device across restarts. Clear history retains saved queries. SQL, names and source paths are stored locally without encryption; result rows and connection credentials are not stored.

Update installation and restart still require an explicit user action. Existing data files, settings and saved workspaces remain compatible. See query history for storage limits and behavior.

All five packages include SHA-256 checksums: macOS Apple Silicon DMG, Windows x64 setup EXE and portable ZIP, and Linux x64 DEB and tar.gz. macOS is ad-hoc signed, not Apple-notarized; Windows packages are unsigned. See the installation guide for platform requirements.

中文

Bamboo 0.2.2 新增查询历史与命名保存查询。

• 编辑器右侧的查询记录按钮打开可搜索弹窗。历史仅记录主动运行且成功的查询；已保存页签保留命名查询，支持重命名与删除。
• 查询仅能打开到原数据源的编辑器，不会自动执行。替换未执行的草稿需要确认，其他数据源的查询无法覆盖当前编辑器。
• 查询历史与已保存查询保存在本机私有配置中，重启后仍可使用。清空历史会保留已保存查询。SQL、名称和数据源路径在本地以未加密形式保存，不保存结果行或连接凭据。

更新安装和重启仍需用户主动确认，现有数据文件、设置和已保存工作区保持兼容。存储限制与操作说明详见查询记录文档。

三个桌面平台同时提供五种包和 SHA-256：macOS Apple Silicon DMG、Windows x64 安装 EXE 与 ZIP、Linux x64 DEB 与 tar.gz。macOS 为临时签名且未经 Apple 公证，Windows 未签名。平台依赖详见安装指南。

Bamboo 0.2.1

Bamboo 0.2.1 improves repeated workspace queries and wide result tables.

• Local Parquet workspace queries can reuse one validated catalog for up to 30 idle seconds. Each query still checks file identity, read access and workspace boundaries; CSV/TSV options continue to bind afresh.
• S3 and R2 workspace queries can reuse verified object ranges while retaining fresh signed HEAD checks, version checks and new scoped gateways. All readers share bounded cache and transfer budgets.
• Wide result pages reuse table cells and avoid expensive automatic column layout. Every column and exact value remains available for scrolling, text selection, browser find and accessibility.

Retained data is not a transactional snapshot of concurrent changes. Cache budgets are not total application memory limits. Existing source files, settings and saved workspaces remain compatible. Update installation and restart still require an explicit user action.

All five packages include SHA-256 checksums: macOS Apple Silicon DMG, Windows x64 setup EXE and portable ZIP, and Linux x64 DEB and tar.gz. macOS is ad-hoc signed, not Apple-notarized; Windows packages are unsigned. See the installation guide for platform requirements.

中文

Bamboo 0.2.1 优化重复工作区查询与宽表结果显示。

• 本地 Parquet 工作区可复用经过验证的目录，空闲 30 秒后释放。每次查询仍检查文件身份、读取权限和工作区范围；CSV/TSV 的读取选项每次重新绑定。
• S3 与 R2 可复用已验证的对象分块，每次仍进行签名 HEAD 和版本检查，并使用新的独立网关。所有读取器共享有界缓存和传输预算。
• 宽表复用单元格并减少自动列宽布局开销，保留全部列、精确值、滚动、文本选择、浏览器查找和辅助功能语义。

缓存不构成并发修改下的事务快照，也不是应用总内存上限。现有文件、设置和已保存工作区保持兼容；更新安装和重启仍需用户主动确认。

三个桌面平台同时提供五种包和 SHA-256：macOS Apple Silicon DMG、Windows x64 安装 EXE 与 ZIP、Linux x64 DEB 与 tar.gz。macOS 为临时签名且未经 Apple 公证，Windows 未签名。平台依赖详见安装指南。

Bamboo 0.2.0

Bamboo 0.2.0 adds complete-result exports, local CSV/TSV input, display controls,
and on-demand analysis while keeping source files read-only.

• Export the complete query: choose the current page or complete result in
  CSV, TSV, JSON or JSONL. Complete exports rerun the last successful query with
  its filter, sorting and projection, stream within resource budgets, show row
  progress and support cancellation. Existing files are never overwritten;
  failed exports are not published as complete files. JSON and JSONL retain
  exact strings, types and NULLs. JSONL starts with a metadata record. CSV/TSV
  protect formula-like text and do not distinguish NULL from empty text.
• Open local CSV and TSV: open single files or discover them in folders,
  restore sessions and join them with Parquet in the same workspace. Choose a
  delimiter, header interpretation and optional conservative type inference.
  UTF-8 is supported, including a BOM, quoted delimiters and multiline fields.
  Text-preserving defaults retain leading zeros and precise numeric text;
  malformed input fails visibly. S3/R2 input remains Parquet-only.
• Adjust display size: General settings provides native application zoom
  from 75% to 200% and data text from 12 to 20 px, with reset and restart
  persistence. SQL text follows the data size with one extra pixel.
• Analyze the complete query: explicitly request exact counts, NULLs,
  distinct values, min/max and the 20 most frequent values. Build a grouped
  COUNT/SUM/AVG/MIN/MAX pivot and export its independent snapshot. Profiles
  reject more than 10,000 distinct non-NULL values; pivots reject more than
  1,000 groups or 4 MiB of value text. These are complete results, never samples
  or silent truncation. AVG follows DuckDB's floating-point SQL semantics.

Complete exports and analysis read sources again; they are not transactional
copies of the original preview. Normal preview limits, read-only SQL validation,
workspace scope checks, exact IPC values, query budgets and cancellation remain.

All three desktop platforms ship together with SHA-256 checksums:

• macOS Apple Silicon, macOS 12+: open
  Bamboo-0.2.0-macOS-arm64.dmg and drag Bamboo into Applications. The app is
  ad-hoc signed, not Apple-notarized; first launch may require approval in
  System Settings → Privacy & Security.
• Windows 10/11 x64: use Bamboo-0.2.0-windows-x64-setup.exe or extract
  Bamboo-0.2.0-windows-x64.zip. Requires Microsoft Edge WebView2 Evergreen x64
  Runtime. Packages are unsigned and may trigger SmartScreen.
• Ubuntu 22.04+ / compatible Debian x64 desktop: install
  Bamboo-0.2.0-linux-x64.deb or extract Bamboo-0.2.0-linux-x64.tar.gz.
  Requires GTK 3, WebKitGTK 4.1, libsecret, libstdc++, glibc 2.35+ and xdg-utils;
  the portable archive does not install them. Other distributions have not
  been validated.

Existing workspaces and open files are restored after restart. macOS update
checks may announce a version; installation and restart still require your action.

中文

Bamboo 0.2.0 新增完整结果导出、本地 CSV/TSV 打开、字号与缩放，以及按需列统计和透视分析。

• 导出可选当前页或完整查询结果，支持 CSV、TSV、JSON、JSONL；完整导出显示进度、
  可取消、不覆盖已有文件，失败时不发布不完整文件。JSON/JSONL 保留精确值和 NULL，
  JSONL 首行为元数据；CSV/TSV 防止文本公式注入，但无法区分 NULL 与空文本。
• 本地文件与文件夹支持 CSV/TSV，保留会话，并可与同工作区的 Parquet 联表查询。
  支持 UTF-8、BOM、引号及多行字段，可调整分隔符、表头和保守类型推断。默认按文本
  读取以保留前导零和精确数字；云端 S3/R2 本轮仍仅支持 Parquet。
• 通用设置可调整 75%–200% 应用缩放、12–20 px 数据字号，支持重置与重启恢复。
• 分析基于完整的最后成功查询，可计算总数、NULL、不同值、最小/最大值与高频值，
  并生成 COUNT/SUM/AVG/MIN/MAX 透视和独立导出。列统计超过 10,000 个不同非 NULL
  值、透视超过 1,000 组或 4 MiB 时明确报错，不采样或静默截断。AVG 遵循 SQL 浮点语义。

完整导出与分析会重新读取源文件，不代表预览时源字节的事务快照。数据只读、路径范围、
资源预算和取消机制继续生效。三个桌面平台同时提供安装包及 SHA-256；macOS 为临时
签名且未经 Apple 公证，Windows 安装包未签名。更新安装与重启仍需用户主动确认。

Bamboo 0.1.5

Bamboo 0.1.5 fixes keyboard focus when sorting query results and adds English and Chinese getting-started documentation.

• Keep the column header focused while a sort is running and after results refresh. Respect focus moved elsewhere, retain the previous sort direction after a failed query, and expose the sort state to assistive technology.
• Document installation on all three platforms, opening local files and S3/R2 workspaces, and read-only SQL examples at getbamboo.app/en/docs/ and getbamboo.app/zh/docs/.
• Expand INT96 compatibility and exact-value regression coverage. Default INT96 reads still use TIMESTAMP with microsecond precision. The documented, existing read_parquet(..., int96_as='struct') option retains nanoseconds for an already-opened, authorized local file; this release does not change the default schema or add a cloud precision setting.

All three desktop platforms ship together with SHA-256 checksums:

• macOS Apple Silicon, macOS 12+: open Bamboo-0.1.5-macOS-arm64.dmg and drag Bamboo into Applications. The app is ad-hoc signed, not Apple-notarized; after the first launch attempt, macOS may require approval in System Settings → Privacy & Security.
• Windows 10/11 x64: use Bamboo-0.1.5-windows-x64-setup.exe, or extract Bamboo-0.1.5-windows-x64.zip. Requires Microsoft Edge WebView2 Evergreen x64 Runtime. Packages are unsigned and may trigger SmartScreen.
• Ubuntu 22.04+ / compatible Debian x64 desktop: install Bamboo-0.1.5-linux-x64.deb, or extract Bamboo-0.1.5-linux-x64.tar.gz. Requires GTK 3, WebKitGTK 4.1, libsecret, libstdc++, glibc 2.35+ and xdg-utils; the portable archive does not install those libraries. Other distributions have not been validated.

Existing workspaces and open files are restored after restart. macOS checks for updates hourly; installation and restart still require your confirmation.

Bamboo 0.1.4

Bamboo 0.1.4 opens wide Parquet files without the previous 256-column preview limit.

• Preview all columns in local files, folder workspaces, Amazon S3 and Cloudflare R2. Wide-file regressions cover up to 4,096 columns.
• Keep previews bounded by reducing rows per page for wide results. For example, a 4,096-column result shows up to 31 rows per page, with a visible explanation and correct next/previous navigation.
• Preserve exact values, NULL and empty text, and support sorting and filtering on later columns. Reading a file does not modify it.
• Retain query resource limits and bound schema metadata. If a schema is too large, select the columns you need in a workspace query.

All three desktop platforms ship together with SHA-256 checksums:

• macOS Apple Silicon, macOS 12+: open Bamboo-0.1.4-macOS-arm64.dmg and drag Bamboo into Applications. The app is ad-hoc signed, not Apple-notarized; after the first launch attempt, macOS may require approval in System Settings → Privacy & Security.
• Windows 10/11 x64: use Bamboo-0.1.4-windows-x64-setup.exe, or extract Bamboo-0.1.4-windows-x64.zip. Requires Microsoft Edge WebView2 Evergreen x64 Runtime. Packages are unsigned and may trigger SmartScreen.
• Ubuntu 22.04+ / compatible Debian x64 desktop: install Bamboo-0.1.4-linux-x64.deb, or extract Bamboo-0.1.4-linux-x64.tar.gz. Requires GTK 3, WebKitGTK 4.1, libsecret, libstdc++, glibc 2.35+ and xdg-utils; the portable archive does not install those libraries. Other distributions have not been validated.

Existing workspaces and open files are restored after restart. macOS checks for updates hourly; installation and restart still require your confirmation.

Bamboo 0.1.3

Bamboo 0.1.3 improves Parquet compatibility and preserves empty records in CSV exports.

• Open SAP-style Parquet DECIMAL columns that omit scale, using the format's default of zero while preserving exact values. Invalid decimal metadata is rejected. Precision above 38 digits is explicitly unsupported instead of being approximated as floating point.
• Keep every single-column CSV record, including leading, middle and trailing NULL or empty text. CSV still represents NULL and empty text alike; JSON keeps them distinct. Nanosecond time values retain their precision.
• Restore open workspaces, independent files and the active selection after restart, with a bounded preview and without automatically running previous SQL.
• Check for macOS updates hourly and show a quiet reminder. Installing and restarting still requires an explicit action.

All three desktop platforms ship together with SHA-256 checksums:

• macOS Apple Silicon, macOS 12+: open Bamboo-0.1.3-macOS-arm64.dmg and drag Bamboo into Applications. The app is ad-hoc signed, not Apple-notarized; after the first launch attempt, macOS may require approval in System Settings → Privacy & Security.
• Windows 10/11 x64: use Bamboo-0.1.3-windows-x64-setup.exe, or extract Bamboo-0.1.3-windows-x64.zip. Requires Microsoft Edge WebView2 Evergreen x64 Runtime. Packages are unsigned and may trigger SmartScreen.
• Ubuntu 22.04+ / compatible Debian x64 desktop: install Bamboo-0.1.3-linux-x64.deb, or extract Bamboo-0.1.3-linux-x64.tar.gz. Requires GTK 3, WebKitGTK 4.1, libsecret, libstdc++, glibc 2.35+ and xdg-utils; the portable archive does not install those libraries. Other distributions have not been validated.

Bamboo 0.1.2

Bamboo 0.1.2 brings the Parquet workbench to macOS Apple Silicon, Windows x64 and Linux x64.

• Open local files and folders, or explicitly configured Amazon S3 and Cloudflare R2 workspaces.
• Explore schemas and exact values, run read-only SQL, and export selected results as CSV or JSON.
• Choose English or Chinese, light or dark appearance, and adjustable memory and temporary-disk budgets.
• Platform-aware downloads and bilingual first-install instructions are available on the website.

This release includes the Windows spill-test verification repair: live file-handle metadata confirms nonempty spill files while DuckDB is writing. Query results, memory limits and cleanup checks remain enforced. Version 0.1.1 was not published; its tag is preserved.

Install

• macOS Apple Silicon, macOS 12+: open Bamboo-0.1.2-macOS-arm64.dmg and drag Bamboo into Applications. The app is ad-hoc signed, not Apple-notarized; after the first launch attempt, macOS may require approval in System Settings → Privacy & Security.
• Windows 10/11 x64: use Bamboo-0.1.2-windows-x64-setup.exe, or extract Bamboo-0.1.2-windows-x64.zip. Requires Microsoft Edge WebView2 Evergreen x64 Runtime. Packages are unsigned and may trigger SmartScreen.
• Ubuntu 22.04+ / compatible Debian x64 desktop: install Bamboo-0.1.2-linux-x64.deb with sudo apt install ./Bamboo-0.1.2-linux-x64.deb, or extract Bamboo-0.1.2-linux-x64.tar.gz. Requires GTK 3, WebKitGTK 4.1, libsecret, libstdc++, glibc 2.35+ and xdg-utils; the portable archive does not install those libraries. Other distributions have not been validated.

Cloud credentials stay in the platform vault: macOS Keychain, Windows Credential Manager or Linux Secret Service. Linux cloud connections require a user D-Bus session and an unlocked Secret Service such as GNOME Keyring.

Every package has a SHA-256 companion file. The signed appcast.xml supplies macOS update checks; installing and restarting require explicit confirmation. Windows/Linux update checks open the releases page for manual installation. No Intel Mac or Windows/Linux ARM64 package is included.

Bamboo 0.1.0

Bamboo's first release brings a focused Parquet workbench to macOS.

• Explore files quickly. Open a Parquet file or folder, inspect its schema, filter and sort rows, and keep multiple workspaces organized.
• Ask questions with SQL. Query a single file or combine files across an opened workspace, then export results as CSV or JSON.
• Read from cloud storage. Connect Amazon S3 and Cloudflare R2 buckets or prefixes with read-only access. Credentials stay in macOS Keychain.
• Work with large datasets. Bounded range reads, reusable caches, and configurable memory and temporary-disk budgets help keep resource use under control.
• Make it yours. Choose English or Chinese, light or dark appearance, or follow the system.
• Stay up to date. See new-version reminders in the sidebar or check in Settings. Installation and restart require your confirmation. Update metadata and downloads are verified with Sparkle signatures and delivered through this repository's GitHub Releases.

Install

Download Bamboo-0.1.0-macOS-arm64.dmg, open it, and drag Bamboo into Applications.

Requires an Apple Silicon Mac running macOS 12 or later. This build is ad-hoc signed and is not Apple-notarized; macOS may require you to allow it in System Settings → Privacy & Security after the first launch attempt.

The accompanying SHA-256 file lets you verify the installer download. The signed appcast.xml is used by the app's automatic update checker.