Sjoerd Mullender wrote:
> On 2008-02-07 18:45, Romulo Goncalves wrote:
>> Update of /cvsroot/monetdb/MonetDB5/src/optimizer
>> In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv22937
>>
>> Modified Files:
>> opt_joinpath.mx
>> Log Message:
>> Revert Martin code because it does not work. The code will remain, but
>> commented
>> for future improvements as Martin suggested...
>
> A better way to comment larger sections of code is to use #if 0/#endif.
#ifdef 0
#endif
does not work:
mizer_la-opt_joinpath.o
/ufs/goncalve/MonetDB/MonetDB5/src/optimizer/opt_joinpath.mx:257:8:
error: macro names must be identifiers
make[5]: *** [lib_optimizer_la-opt_joinpath.lo] Error 1
make[5]: Leaving directory
`/export/scratch1/goncalve/MonetDB/debug/MonetDB5/build/src/optimizer'
make[4]: *** [all] Error 2
make[4]: Leaving directory
`/export/scratch1/goncalve/MonetDB/debug/MonetDB5/build/src/optimizer'
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory
`/export/scratch1/goncalve/MonetDB/debug/MonetDB5/build/src'
make[2]: *** [all] Error 2
make[2]: Leaving directory
`/export/scratch1/goncalve/MonetDB/debug/MonetDB5/build/src'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory
`/export/scratch1/goncalve/MonetDB/debug/MonetDB5/build'
make: *** [all] Error 2
[EMAIL PROTECTED] build]$ vim
~/MonetDB/MonetDB5/src/optimizer/opt_joinpath.mx
Romulo
> That nests.
>
>>
>>
>> Index: opt_joinpath.mx
>> ===================================================================
>> RCS file: /cvsroot/monetdb/MonetDB5/src/optimizer/opt_joinpath.mx,v
>> retrieving revision 1.21
>> retrieving revision 1.22
>> diff -u -d -r1.21 -r1.22
>> --- opt_joinpath.mx 7 Feb 2008 16:57:18 -0000 1.21
>> +++ opt_joinpath.mx 7 Feb 2008 17:44:59 -0000 1.22
>> @@ -231,22 +231,30 @@
>> The join path optimizer takes a join sequence and
>> attempts to minimize the intermediate result.
>> The choice depends on a good estimate of intermediate
>> -results.SAMPLE_THRESHOLD_lOG
>> +results.
>> For the time being, we use a simplistic model, based
>> on the assumption that most joins are foreign key joins anyway.
>> +
>> We use a sample based approach for sizeable tables.
>> -The model is derived from the select statement.
>> +The model is derived from the select statement. However, we did not succeed.
>> +The code is now commented for future improvement.
>> +Final conclusion from this exercise is:
>> +The difference between the join input size and the join output size is not
>> +the correct (or unique) metric which should be used to decide which order
>> +should be used in the joinPath.
>> +
>> @c
>> -#define SAMPLE_THRESHOLD_lOG 17
>> +/*#define SAMPLE_THRESHOLD_lOG 17*/
>> size_t
>> ALGjoinCost(BAT *l, BAT *r)
>> {
>> - size_t lc, rc, lsize,rsize;
>> - BAT *lsample, *rsample, *j;
>> + size_t lc, rc;
>> +/* size_t lsize,rsize;
>> + BAT *lsample, *rsample, *j; */
>>
>> lc = BATcount(l);
>> rc = BATcount(r);
>> - if( lc > 100000 || rc > 100000){
>> +/* if( lc > 100000 || rc > 100000){
>> lsize= MIN(lc/100, (1<<SAMPLE_THRESHOLD_lOG)/3);
>> lsample= BATsample(l,lsize);
>> BBPreclaim(lsample);
>> @@ -257,7 +265,7 @@
>> lsize= BATcount(j);
>> BBPreclaim(j);
>> return lsize;
>> - }
>> + }*/
>> if( l->ttype== TYPE_oid || r->htype== TYPE_oid )
>> return MIN(lc,rc);
>> return lc * rc;
>>
>>
>> -------------------------------------------------------------------------
>> This SF.net email is sponsored by: Microsoft
>> Defy all challenges. Microsoft(R) Visual Studio 2008.
>> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>> _______________________________________________
>> Monetdb-checkins mailing list
>> [EMAIL PROTECTED]
>> https://lists.sourceforge.net/lists/listinfo/monetdb-checkins
>
>
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Monetdb-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-developers
