Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I wrote a new, general-purpose data management system that was originally designed to compete with file systems for storing large amounts of unstructured data (my background is in disk utilities like PartitionMagic and Drive Image). It can be used to convert 100M+ files (photos, documents, videos, etc.) into data objects and attach searchable, contextual meta-data tags to them.

The tags I invented were essentially highly-optimized key-value stores that worked really well. I then discovered that these same KV stores could also be used to form relational tables (a columnar store). It could be used to store both highly structured data or the more semi-structured data found in things like Json documents where every value could be an array. Queries were lightning fast and it could perform analytic operations while still performing transactions at a high speed.

My hobby turned into a massive project as I figured out how to import/export CSV, Json, Json lines, XML, etc. files to/from tables easily and quickly. It still has a lot of work to go before it is a 'full-blown' database, but it is now a great tool for cleaning and analyzing some pretty big tables (tested to 2500 columns and 200M rows).

It is in open beta at https://didgets.com/download and there are a bunch of short videos that show some of the things it can do on my youtube channel: https://www.youtube.com/channel/UC-L1oTcH0ocMXShifCt4JQQ



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: