Re: [sqlite] indexing speed

2010-08-02 Thread Alexey Pechnikov
I think you may use FTS3 extension. This is very fast for big tables too. I did test about to 500 millions rows with UUIDs in FTS3 table and results are fine. See as example test here: http://book.mobigroup.ru/dir?name=web_project_DBMS/distributed_schema Some result logs you can find in files perft

Re: [sqlite] indexing speed

2010-08-02 Thread Simon Slavin
On 2 Aug 2010, at 5:31pm, Paul Sanderson wrote: > I have a table with just a few columns, one is a SHA1 hash and the > second an MD5, there are about 17 Million rows in the table > > if I create an index on the SHA1 column using "create index if not > exists sha1index on hashtable(sha1)" the pro

Re: [sqlite] indexing speed

2010-08-02 Thread Igor Tandetnik
Paul Sanderson wrote: > I have a table with just a few columns, one is a SHA1 hash and the > second an MD5, there are about 17 Million rows in the table > > if I create an index on the SHA1 column using "create index if not > exists sha1index on hashtable(sha1)" the process takes about 3 > minute

[sqlite] indexing speed

2010-08-02 Thread Paul Sanderson
I have a table with just a few columns, one is a SHA1 hash and the second an MD5, there are about 17 Million rows in the table if I create an index on the SHA1 column using "create index if not exists sha1index on hashtable(sha1)" the process takes about 3 minutes, if I follow this immediately by