We are trying to improve the query performance of our virtual table
implementation (which is implemented in C). Due to requirements of
external code, a specified column of a specified row (by rowid) is
queried at a time (the least efficient means of access). Our virtual
table is accessing entries in a memory-based array.
I have implemented a benchmark script written in Python using the APSW
wrapper. The benchmark script reveals that access to a native
database table is 5 times faster than access to our virtual table.
Intuitively, I would think that access to a memory-based virtual table
could be faster than native tables. Our developer has implemented
xBestIndex and xFilter support which is intended to result in direct
access to the requested row rather than scanning the whole table.
What is the expected performance of a properly implemented virtual
table (assuming little additional overhead) vs a native table?
Bob
--
Bob Friesenhahn
bfrie...@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer, http://www.GraphicsMagick.org/
_______________________________________________
sqlite-users mailing list
sqlite-users@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
