Re: [sqlite] Is there a way to create a foreign key in existing table?

2013-01-01 Thread Peter Haworth
And remembering that you will lose any indexes or triggers defined for the table by doing this. There are plenty of third party tools out there that will take care of adding/removing constraints to existing tables and a whole lot of other functions that aren't available in SQLite's DDL, while acco

Re: [sqlite] Write performance question for 3.7.15

2013-01-01 Thread Dan Frankowski
Ah, interesting. However, yes, we need production-ready. Good luck with sqlite4 tho. On Tue, Jan 1, 2013 at 11:43 AM, Richard Hipp wrote: > On Tue, Jan 1, 2013 at 12:33 PM, Dan Frankowski > wrote: > > > > > We are comparing to leveldb, which seems to have much better write > > performance even

Re: [sqlite] Write performance question for 3.7.15

2013-01-01 Thread Richard Hipp
On Tue, Jan 1, 2013 at 12:33 PM, Dan Frankowski wrote: > > We are comparing to leveldb, which seems to have much better write > performance even in a limited-memory situation. Of course it offers much > less than sqlite. It is a partially-ordered key/value store, rather than a > relational databa

Re: [sqlite] Write performance question for 3.7.15

2013-01-01 Thread Dan Frankowski
I appreciate everyone's thoughts about this. Knowing larger batch sizes help is interesting. Unfortunately, we don't always control the batch size. We're using 1000 as an optimistic estimate, but we receive things and may just have to commit after awhile. Knowing that more OS file cache or a fast

Re: [sqlite] SQLite4 Performance

2013-01-01 Thread Michael Black
Thanks...progress...now we get a SIGBUS later on... Program received signal SIGBUS, Bus error. 0x0046ce4c in treeShmalloc (pDb=0x6c9728, bAlign=0, nByte=25, pRc=0x7fffd784) at src/lsm_tree.c:668 668 pNext->iNext = 0; pRc looks suspiciously like it's ready to overflow #inc

Re: [sqlite] How to verify referential integrity of SQLite database

2013-01-01 Thread Richard Hipp
On Thu, Dec 27, 2012 at 11:43 AM, Levi Haskell (BLOOMBERG/ 731 LEXIN) < lhask...@bloomberg.net> wrote: > Hi Richard, > > Sorry, I missed your message before. > > Look, in my example: > > SQLite version 3.7.16 2012-12-20 01:15:20 > Enter ".help" for instructions > Enter SQL statements terminated wi

Re: [sqlite] SQLite4 Performance

2013-01-01 Thread Dan Kennedy
On 01/31/2013 11:13 PM, Michael Black wrote: Do we still get to report bugs? I checked out the fossil repository fossil clone http://www.sqlite.org/src4 sqlite4.fossil I tried my insert test and ran into a problem. I'm running Redhat 5.7 gcc 4.4.4 This program dies (showing inserts/sec) 123536