Hi Karen,
i didn't want to bother you too much (because of that I dindnt send too much
code).
I think these model are bad implemented. I have put the raw_id_admin
parameter to true but the point is that I have too many queries:
>>> len(connection.queries)
53
and some of them, all that goes to sportsound_plantilla_metodo_embebido, are
really slow (around 1.7 seconds):
{'time': '1.788', 'sql': u'SELECT
`sportsound_plantilla_metodo_embebido`.`id`,`sportsound_plantilla_metodo_embebido`.`orden`,`sportsound_plantilla_metodo_embebido`.`ejercicio_id`,`sportsound_plantilla_metodo_embebido`.`numero_repeticiones`,`sportsound_plantilla_metodo_embebido`.`repeticiones_maximas`,`sportsound_plantilla_metodo_embebido`.`hasta_extenuacion`,`sportsound_plantilla_metodo_embebido`.`distancia`,`sportsound_plantilla_metodo_embebido`.`tiempo_ejecucion_m`,`sportsound_plantilla_metodo_embebido`.`tiempo_ejecucion_sg`,`sportsound_plantilla_metodo_embebido`.`altura`,`sportsound_plantilla_metodo_embebido`.`velocidad_excentrica`,`sportsound_plantilla_metodo_embebido`.`velocidad_concentrica`,`sportsound_plantilla_metodo_embebido`.`tipo_contraccion`,`sportsound_plantilla_metodo_embebido`.`ritmo_carrera`,`sportsound_plantilla_metodo_embebido`.`unidad_ritmo_carrera`,`sportsound_plantilla_metodo_embebido`.`frecuencia_cardiaca`,`sportsound_plantilla_metodo_embebido`.`frecuencia_cardiaca_1m`,`sportsound_plantilla_metodo_embebido`.`frecuencia_cardiaca_3m`,`sportsound_plantilla_metodo_embebido`.`porcentaje_carga`,`sportsound_plantilla_metodo_embebido`.`intensidad_subjetiva`,`sportsound_plantilla_metodo_embebido`.`peso`,`sportsound_plantilla_metodo_embebido`.`evaluar_numero_repeticiones`,`sportsound_plantilla_metodo_embebido`.`evaluar_distancia`,`sportsound_plantilla_metodo_embebido`.`evaluar_tiempo_ejecucion`,`sportsound_plantilla_metodo_embebido`.`evaluar_altura`,`sportsound_plantilla_metodo_embebido`.`evaluar_peso`,`sportsound_plantilla_metodo_embebido`.`evaluar_ritmo_carrera`,`sportsound_plantilla_metodo_embebido`.`evaluar_frecuencia_cardiaca`,`sportsound_plantilla_metodo_embebido`.`evaluar_frecuencia_cardiaca_1m`,`sportsound_plantilla_metodo_embebido`.`evaluar_frecuencia_cardiaca_3m`,`sportsound_plantilla_metodo_embebido`.`evaluar_intensidad`
FROM `sportsound_plantilla_metodo_embebido`'}
Actually these queries select all the parameters, that doesnt make any sense
( maybe is the behaviour of manytomanyfield)
Thank you very much karen for your patient and support.
tomorrow it will be a better day :)
Miguel
Sent from Madrid, Spain
On Wed, Apr 1, 2009 at 8:33 PM, Karen Tracey <kmtra...@gmail.com> wrote:
> On Wed, Apr 1, 2009 at 1:35 PM, Miguel <migue...@gmail.com> wrote:
>
>> sorry karen. I have tried both. First I put in teh forma_parte_de foreing
>> key field but it has no discernible effect on performance. I even execute it
>> via python console and you can see clearly that it takes long time to
>> response.
>>
>> Then I tried to put the raw parameter in the many to many fields but I
>> observe no effect .
>>
>> Maybe, the more of 100000 entries of Plantilla_metodo_embebido are the
>> responsible of the long time ...
>>
>
> Well you still haven't shared much beyond little snippets of the model
> involved here so it's getting a little hard to know what else to suggest.
> The number of rows in the related table doesn't matter if you specify
> raw_id_admin=True. Consider this model from my own app:
>
> class Clues(models.Model):
> ID = models.AutoField(primary_key=True)
> Theme = models.CharField(maxlength=3, choices=YESNO_CHOICES,
> verbose_name='Theme?', core=True)
> EntryID = models.ForeignKey(Entries, verbose_name='Entry', db_column =
> 'Entry ID', edit_inline=models.TABULAR, num_in_admin=3, core=True,
> raw_id_admin=True)
> PuzzleID = models.ForeignKey(Puzzles, verbose_name='Puzzle', db_column
> = 'Puzzle ID', edit_inline=models.TABULAR, num_in_admin=3,
> raw_id_admin=True)
> Clue = models.CharField(maxlength=150, core=True)
> Num = models.IntegerField(null=True, blank=True, core=True)
> Dir = models.CharField(maxlength=1, core=True)
> Derived = models.BooleanField()
> [remainder deleted]
>
> If I remove the raw_id_admin=True from the EntryID (table with approx.
> 130,000 rows) and PuzzleID (table with approx. 10,000 rows) fields and try
> to create a ChangeManipulator for this model (which itself has approx.
> 830,000 rows in the DB) from a shell, it takes about 30 seconds. If I have
> DEBUG set to True I can use django.db.connection.queries to see where much
> of that time is coming from:
>
> Python 2.5.2 (r252:60911, Feb 21 2008, 13:11:45) [MSC v.1310 32 bit
> (Intel)] on win32
> Type "help", "copyright", "credits" or "license" for more information.
> (InteractiveConsole)
> >>> from crossword.models import Clues
> >>> from django.db import connection
> >>> x = Clues.ChangeManipulator(2432)
> >>> len(connection.queries)
> 3
> >>> connection.queries[0]
> {'time': '0.000', 'sql': 'SELECT
> `Clues`.`ID`,`Clues`.`Theme`,`Clues`.`Entry ID`,`Clues`.`Puzzle
> ID`,`Clues`.`Clue`,`Clues`.`Num`,`Clues`.`Dir`,`Clues`.`Derived` FROM
> `Clues` WHERE (`Clues`.`ID` = 2432)'}
> >>> connection.queries[1]
> {'time': '13.560', 'sql': 'SELECT `Entries`.`Entry
> ID`,`Entries`.`Entry`,`Entries`.`Exclude` FROM `Entries` ORDER BY
> `Entries`.`Entry` ASC'}
> >>> connection.queries[2]
> {'time': '0.901', 'sql': 'SELECT `Puzzles`.`Puzzle ID`, [remainder deleted
> for readability]}
> >>> quit()
>
> So the unrestricted query on the >130,000 row Entries table needed to
> contruct the drop down list for the select box for the EntryID field takes
> nearly 14 seconds all by itself (this is not my actual production DB, it's
> an old test copy on an old Windows box, perfect for exacerbating performance
> issues). The corresponding query for the Puzzles table (an order of
> magnitude smaller) only takes nearly a second. Apparently the processing
> required to massage those results into the proper format within the
> ChangeManipulator means the actual time to create the change manipulator is
> about double the time required for the queries.
>
> If I restore the raw_id_admin=True for both those fields, creating the
> change manipulator becomes nearly instantaneous, as those expensive queries
> are no longer needed:
>
> Python 2.5.2 (r252:60911, Feb 21 2008, 13:11:45) [MSC v.1310 32 bit
> (Intel)] on win32
> Type "help", "copyright", "credits" or "license" for more information.
> (InteractiveConsole)
> >>> from crossword.models import Clues
> >>> from django.db import connection
> >>> x = Clues.ChangeManipulator(8232)
> >>> len(connection.queries)
> 1
> >>> connection.queries[0]
> {'time': '0.051', 'sql': 'SELECT
> `Clues`.`ID`,`Clues`.`Theme`,`Clues`.`Entry ID`,`Clues`.`Puzzle
> ID`,`Clues`.`Clue`,`Clues`.`Num`,`Clues`.`Dir`,`Clues`.`Derived` FROM
> `Clues` WHERE (`Clues`.`ID` = 8232)'}
> >>>
>
> Perhaps you could look at what queries are being generated in your case --
> that might give you a clue where the time is being spent in your case.
>
> Karen
>
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Django users" group.
To post to this group, send email to django-users@googlegroups.com
To unsubscribe from this group, send email to
django-users+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/django-users?hl=en
-~----------~----~----~----~------~----~------~--~---
