One of my first developer projects. Was working for a GIS company reverse engineering ancient geospatial databases (and our own). Nothing had foreign keys, there was a "global connection table" that provided many-many-joins, primary keys were strings, etc. Hundreds of tables and thousands of columns.
Most of our days on the data migration team were spent tracking down missing connections between entities, inconsistencies in the data that our clients had found, etc. I took a "find a string in every column" stored procedure and rewrote it as a python desktop app. Once the basic functionality was there I multithreaded it to run each query simultaneously. After that, I provided a graph-layout GUI that the users could click through and build their queries dynamically based on existing connections they'd already found.
No idea if it's still a thing anymore but it was amazing for my team. We went from guessing at relationships between entities because we didn't want to wait for the stored proc to run for hours to knowing which entities had what relationships in minutes. Single best thing I've ever written.
Most of our days on the data migration team were spent tracking down missing connections between entities, inconsistencies in the data that our clients had found, etc. I took a "find a string in every column" stored procedure and rewrote it as a python desktop app. Once the basic functionality was there I multithreaded it to run each query simultaneously. After that, I provided a graph-layout GUI that the users could click through and build their queries dynamically based on existing connections they'd already found.
No idea if it's still a thing anymore but it was amazing for my team. We went from guessing at relationships between entities because we didn't want to wait for the stored proc to run for hours to knowing which entities had what relationships in minutes. Single best thing I've ever written.