Re: [sqlite] TCL Test failures on ARM

On 03/12/2013 10:48 AM, bkk wrote: Has anyone came across same problem ? Appreciate if someone can point out what’s going wrong with these tests Is there something odd about trying to write to errno on this platform? These tests depend on function tsIsFailErrno() in file test_syscall.c being

Re: [sqlite] TCL Test failures on ARM

Has anyone came across same problem ? Appreciate if someone can point out what’s going wrong with these tests -- View this message in context: http://sqlite.1065341.n5.nabble.com/TCL-Test-failures-on-ARM-tp67612p67637.html Sent from the SQLite mailing list archive at Nabble.com.

Re: [sqlite] Fuzzy joins

Yeah, one benefit of a long commute is that I could puzzle it out and understand why it works. select t1.*, (  select v from t2 where t2.k1=t1.k1 and t2.k2=t1.k2 and t2.k3=t1.k3 and t2.k4=t1.k4 and t2.k5=t1.k5  group by k1,k2,k3,k4,k5  having count(*)=1  union  select v from t2 where

Re: [sqlite] Fuzzy joins

On Mar 11, 2013, at 10:32 PM, David Bicking wrote: > Um, I am wrong, cause I just tried it and sqlite only returns the level 5 > result. I have no clue why! The joins are setup from most specific to broadest. Each join is evaluated only if the previous one returns null

Re: [sqlite] Fuzzy joins

Ouch, I certainly did forget the where clauses! Table1 and Table2 are going to be roughly the same size. What I am trying to do, btw, is to take an extract from a prior period, which has had the "Value" added to it, and move that value in to the matching record in the new extract. There is a

Re: [sqlite] Fuzzy joins

On Mar 11, 2013, at 4:54 PM, David Bicking wrote: > Am I missing an obviously better way to do it? > A way that can easily be expanded when they come back to me and say if I > looked at a fifth column, you'd have been able to match it…. As they stand, your updates will

Re: [sqlite] SQLite 3.7.16 beta

On Mon, Mar 11, 2013 at 11:48 AM, Richard Hipp wrote: > > > On Sat, Mar 2, 2013 at 11:29 AM, Richard Hipp wrote: > > > > Please report any problems encountered. > > > > -- > > D. Richard Hipp > > d...@sqlite.org Some lunchtime reading resulted in this

Re: [sqlite] Changed behavior in SQLite 3.7.16 beta

OK, downloaded and built from this and the problem went away, including the more complex case the submitted case was cut down from.  We'll run our autotests again and see if anything else pops up.  Thanks for all the hard work! Peter - Original Message > From: Richard Hipp

Re: [sqlite] Windows-Specific 2-c-files Amalgamation?

Can't you just wait to fix things until "patch Tuesday"??? :-) Sheesh...we can't handle things this quickly... -Original Message- From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of Richard Hipp Sent: Monday, March 11, 2013 12:37 PM To: General

Re: [sqlite] Windows-Specific 2-c-files Amalgamation?

On Mon, Mar 11, 2013 at 1:33 PM, Michael Black wrote: > Richard...he split is missing the 5 split files. I thought the small size > was suspicious. > Fixed. -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list

Re: [sqlite] Windows-Specific 2-c-files Amalgamation?

Richard...he split is missing the 5 split files. I thought the small size was suspicious. -Original Message- From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of Richard Hipp Sent: Monday, March 11, 2013 11:42 AM To: General Discussion of SQLite

Re: [sqlite] Windows-Specific 2-c-files Amalgamation?

On Mon, Mar 11, 2013 at 12:17 PM, Dominique Devienne wrote: > Perhaps someone knows > a trick or two to work-around that MS debugging issue? > Run: "make sqlite3-all.c" to build an amalgamation composed of smaller files. See the http://www.sqlite.org/draft/download.html

[sqlite] Windows-Specific 2-c-files Amalgamation?

I understand the "deployment" ease and performance advantage of the amalgamation. I like it. Except I can't debug it... 'Cause I'm primarily on Windows, which has a well-known limitation in the way it stores Debug Information (uses 16-bit integer for the line number, so limited to ~64K lines,

Re: [sqlite] SQLite 3.7.16 beta

On 11 Mar 2013, at 3:48pm, Richard Hipp wrote: > http://www.sqlite.org/draft/download.html I'm afraid all I can help with is point out a grammatical problem: "and templates (2) is used" I think should have 'templates' in the singular. I add a note that release 3.7.4 was

[sqlite] Fuzzy joins

This is a weird request. I have a table of data with no natural primary key. I need to update this table from a prior table, but some of the data fields can change over time, so I must be flexible on how I match. So the matching I need to do is something like this, if Key1 is unique in both

Re: [sqlite] SQLite 3.7.16 beta

Thanks, everyone, for your help in testing 3.7.16. A number of problems have been found and fixed. Because of these problems, the anticipated release of 3.7.16 has been delayed until one week from today (2013-03-18). The draft website, and especially the amalgamation snapshot on the

Re: [sqlite] per-context resources for user-defined function

On Mon, 11 Mar 2013 12:47:02 +0100 Clemens Ladisch wrote: > James K. Lowden wrote: > > I'm not sure how to manage the lifetime of ancillary data for a > > user-defined function added by sqlite3_create_function (). > > http://www.sqlite.org/c3ref/aggregate_context.html > > >

Re: [sqlite] TCL Test failures on ARM

Any one knows how to fix it ? -- View this message in context: http://sqlite.1065341.n5.nabble.com/TCL-Test-failures-on-ARM-tp67612p67621.html Sent from the SQLite mailing list archive at Nabble.com. ___ sqlite-users mailing list

Re: [sqlite] per-context resources for user-defined function

On 11/03/2013 7:47 AM, Clemens Ladisch wrote: James K. Lowden wrote: I'm not sure how to manage the lifetime of ancillary data for a user-defined function added by sqlite3_create_function (). [...] Suppose xStep doesn't fail, but another query is executing simultaneously, also using the

Re: [sqlite] per-context resources for user-defined function

On Mon, Mar 11, 2013 at 12:52 PM, James K. Lowden wrote: > I need one data structure > per context. I don't see support for that in the SQLite API. Am I > expected to maintain my own global associative array? Not if you use sqlite3_aggregate_context(). See

Re: [sqlite] per-context resources for user-defined function

James K. Lowden wrote: > I'm not sure how to manage the lifetime of ancillary data for a > user-defined function added by sqlite3_create_function (). > [...] > Suppose xStep doesn't fail, but another query is executing > simultaneously, also using the median() UDF. I need one data structure > per

[sqlite] per-context resources for user-defined function

Hello all, I'm not sure how to manage the lifetime of ancillary data for a user-defined function added by sqlite3_create_function (). Consider a median() function. The data are captured on each xStep, then sorted and the median produced in xFinal. Suppose sqlite3_realloc() fails during