[ZPT] Re: Fwd: Re: sql not that persistent

[Michal Kurowski]

Dieter Maurer [dieter@handshake.de] wrote:
> Michal Kurowski writes:
>  > ...
>  >  sql method iterated on five tables ( some scripts involved etc)
>  >                 |         
>  > 		\|/
>  >  zpt: tal:repeat on each of the tables
>  > 		 |
>  > 		\|/
>  >  zpt: tal:repeat on rows from those tables
>  > 		 |
>  > 		\|/
>  >  zpt: link to a script
>  > 		 |
>  > 		\|/
>  >  script: performing an action on fields from that second iteration
>  > 	  and trying to access values from the first iteration
>  > 
>  > As I said: I can access the values (from the "second iteration") 
>  > directly through parameters but I still have to do some sql 
>  > in the script later on.
> Z SQL results live as long as they are referenced.
> They disappear at least when the request is finished.
> 
> You can bind them to names:
> 
>     <dtml-let my_result=sql_result>...</dtml-let> (DTML)
> 
> or
> 
>     <... tal:define="[global] my_result here/sql_result">... (ZPT)
> 
> or put them into the request object
> 
>    "REQUEST.set('my_result', sql_result())" (DTML, similar for ZPT).

Yes, and that's the place that is a bit mysterious to me.

Let's say I have a  global "table" variable defined in ZPT that points
to "here.get_db()".

When I try:

tal:content="python: request.set('my_table', table)"

or

tal:content="python: request.set('my_table', here.get_db())

and then, in a script:

for key, item in request.items():
    print ...

there's no such thing set :-(

Obviously, I am missing something major in here.

Isn't that only strings can be placed in those session/request
dictionaties ?


-- 
Michal Kurowski
<michal@poczta.gazeta.pl>