DSL for table data analysis — with 2-click .exe export. No Python. No PyInstaller. No dependency hell.

ArrayVator is a domain-specific language for analyzing table data. You write in a readable syntax and get results like from SQL + Python + Excel — without needing to know all three.

Key ideas:

• 📊 Table operations — matrices, vectors, slices
• 🚀 Two engines — Matrix (RAM) and DuckDB (Big Data)
• 🧠 Smart help — autocomplete + built-in help panel
• 💬 Clear errors — with hints on how to fix
• ⚡ 2-click .exe export — package your script into a standalone executable

Write your script → press Tools → Export to EXE → get a standalone .exe.

That's it. No Python, no PyInstaller, no .spec files, no DLL hunting.

1. Save your code as .arrv in the editor (Ctrl+S)
2. Open Tools → Export to EXE — the compiler window opens
3. In the compiler:
   • Source .arrv: — choose your saved .arrv file (Browse...)
   • Output folder: — choose where to save the .exe (Browse...)
   • EXE name: — type a name (latin letters, no .exe suffix)
4. Click Build EXE (green button)
5. Done! Your standalone .exe is ready

💡 First time? Click Rebuild Runtime first — the compiler will prepare the runtime files. After that, builds take seconds.

If you've ever tried to package a Python script into a .exe, you know the pain:

• pip install pyinstaller
• Write a .spec file
• Handle missing modules with --hidden-import
• Copy DLLs manually
• Debug mysterious errors like ModuleNotFoundError in frozen builds
• Hope it works on the target machine

ArrayVator does all of this for you.

The compiler scans your code and picks the right runtime:

• runner_minimal — pure language (fastest, smallest)
• runner_data — with Excel / CSV / DuckDB
• runner — full: charts + HTML reports

• ⚡ Build .exe in 2 clicks — no Python, no PyInstaller
• 📊 Matrices and vectors — tables with headers, slicing, ranges
• 🔍 Filtering — filterif, deleteif vs classic loops
• 📈 Aggregates — sum, avg, min, max, count, median, std
• 🎯 Conditional aggregates — sumif, countif, avgif, minif, ...
• 📊 Grouping — groupby, pivot, groupagg
• 🪟 Window functions — rownumber, rank, lag, lead, winsum, qualify
• 📅 Dates and time — year, adddays, DateDiff, datetrunc
• 🧮 Strings — split, joinvector, replacetext, clean, trim
• 🔎 Search — find, vlookup
• 🎨 Conditions — case, applyif
• 🚀 BigData — up to 1 billion rows via DuckDB
• 📉 Charts — Matplotlib / Plotly (bar, line, pie, hist, scatter, box, heatmap, pair)
• 📄 HTML reports — with interactive Plotly charts
• 📁 Files — Excel, CSV, TXT, Parquet, SQLite
• 🧠 Built-in help — click on a function → help appears on the right

Get the latest portable build: Releases

Windows — ArrayVator_Portable.zip:

• ArrayVator.exe — editor (double-click to run)
• runner/ — full interpreter (DuckDB, Excel, charts)
• runner_data/ — data mode (DuckDB, Excel, no charts)
• runner_minimal/ — minimal (core only)
• compiler/ — EXE compiler (GUI)
• help/ — help files (Russian + English)

No Python required! Works on Windows 10/11 (x64).

Installation:

1. Download ArrayVator_Portable.zip
2. Extract anywhere
3. Run ArrayVator.exe

ArrayVator is currently unsigned. Windows SmartScreen may warn:

Windows protected your PC

This is normal. Click More info → Run anyway.

The source is 100% open — verify on GitHub, or build from source.

We're applying for a free open-source code-signing certificate (via SignPath.io). Future releases will be signed.

m = ["Name", "Department", "Salary";
     "Anna", "IT", 85000;
     "Bob", "HR", 65000;
     "Eve", "IT", 95000]

r = filterif(m[:, "Department"] == "IT")
print(r)

Output:

Name  Department  Salary
Anna  IT          85000
Eve   IT          95000

The examples/ folder contains 68 self-contained recipes organized by topic:

Each example is a self-contained file with comments, code, and expected output. Run them with F5 in the editor.

Requirements:

• Python 3.10+
• Windows / Linux / macOS

Setup:

git clone https://github.com/Arrayvator/arrayvator.git
cd arrayvator
pip install -r requirements.txt
python main.py

Dependencies: duckdb, python-calamine, pyexcelerate, openpyxl, matplotlib, plotly, playwright, pyinstaller.

• examples/ — 68 ready-to-run recipes
• help/ — help files (Russian + English)

Some key rules of ArrayVator:

1. All functions return a value — save the result:

   m = filterif(m[:, "Department"] == "IT")
   

2. None is the only empty-value literal (null, nan are forbidden).
3. = — assignment. == — comparison.
4. Logical operators: and, or, not.
5. Indexing starts at 1 — m[1, :] is the header row.
6. Header rule:
   • m[:, "Name"] — header does NOT take part
   • m[:, 3] — header DOES take part
7. DuckDB (BigData) is read-only — some functions are Matrix-only.

See examples/ for details.

MIT — see LICENSE.

Arrayvator

• Website: arrayvator.ru
• GitHub: @Arrayvator