Database, libgist

Because of work needs, I re-studied the database technology. My practice is to start searching the Internet from the algorithm, and the algorithms of the database technology itself have not changed much. I used hash, btree, 10 years ago, these are the things now !! Algorithms are the most stable and slowest part of IT evolution.

 

I haven't used commercial relational databases for a long time. I am naturally tired of SQL. Maybe I wrote too many SQL statements 10 years ago, so I am exhausted when I saw this.

 

After searching on the Internet for a long time, everything has not changed. The open source databases are MySQL and SQlite, and there is no other climate. Memory databases are still not widely used, and FastDB is still fighting against open-source memory databases.

 

After downloading the FastDB code, I suddenly paid attention to an optional index library of FastDB, libgist. I used to know the existence of this library, but I did not read it carefully. I am interested in it this time, download the study for a while.

 

I don't know, I don't know. libgist is a good thing. Some time ago, I put GDBM together and put it in a file to compile VC9. There is no problem in using GDBM in Windows, however, GDBM (similar to berklyDB, TC) features are still weak. I am not happy, And libgist immediately gave me hope.

 

Libgist is actually quite useful. Building Btree and Rtree indexes is much more powerful than Hash-based GDBM (bdb also has btree, But I think libgist is more direct to use, and this library is smaller than bdb)

 

Libgist is earlier than fastdb. I know that the Russian who wrote fastdb also seems to have a reference object. He didn't build a car behind closed doors.

 

We plan to encapsulate libgist in squirrel to create a fast data storage and retrieval function !!