Recorded DB problem

[peter.ludemann]

@AlexeyKosarchuk

So far that list contains around 300-500 terms, but in future it will be much, much more. I expect to run into troubles when scaling data up.

I have some code that needs to store 10s of thousands of terms (a global symbol table). After a few experiments, I ended up using library(rbtrees). (I didn’t try @EricGT 's suggestion of N+K trees.) I also wrote some custom code to read and write them quickly, using library(fastrw).