I invented a data structure with two roots and shared leaves — the Conjoined Tree

Hi folks, I recently invented and open-sourced a new data structure called the Conjoined Tree.

It’s designed for multi-dimensional data modeling—think pivot tables, OLAP cubes, or feature stores—where operations like sorting, filtering, or inserting can be done along any axis, without needing to shift the whole structure.

Instead of one root like a traditional tree, it has two or more root nodes (like rows and columns), and all share the same leaf nodes. It scales up to N dimensions and supports sparse data naturally.

🔗 GitHub: https://github.com/esuryadi/conjoined-tree

📝 Article: https://www.linkedin.com/pulse/tree-two-roots-new-data-structure-multi-dimensional-thinking-suryadi-czycc

📚 Spec: https://edsuryadi.atlassian.net/wiki/external/NzM1Mjk1NDUwOWU4NGIwZDg5MGQxYWMzMTdhN2M4YWM

Would love feedback or collaboration ideas. Feel free to contribute or build something with it!