Hi Henning,
I do not have code yet. I have just done a quick survey to see whether
there are any big obstacles, but I did not find any. I just wanted to make
sure the work is done only once. So if you already have code, just tell me
and go ahead.
Thomas
"Henning P. Schmiedehausen" <[EMAIL PROTECTED]> schrieb am 10.12.2004
12:57:47:
> Thomas Fischer <[EMAIL PROTECTED]> writes:
>
> >Hi,
>
> >In the org.apache.torque.util.SQLBuilder class (recently taken out of
> >BasePeer), there is a lot of code which parses column names in lots of
> >places. I would like to go ahead and see if the existing code can be
> >simplified by creating an internal class in SQLBuilder which represents
a
> >database (pseudo)column, does the parsing and takes care of things like
> >aliases. I am not sure whether this works out, as there might be subtle
> >differences in the parsing of raw column names e.g. between the FROM
part
> >and the ORDER BY part of the query, but I'd like to give it a try.
>
> Ah, coincidence strikes again. Part of my reworks was to get the code
> in shape to do exactly that. So yes, this is where I will work right
> now.
>
> >While doing this, I would also like to check whether static schema
support
> >can be easily built in. My fist step would be to allow tablenames which
> >contain a dot (like SCHEMANAME.TABLENAME). This would mean to replace
the
> >splitting between tablename and columnname, which at the moment looks at
> >the first dot in the column name (rawColumnName.indexOf('.')), by
splitting
> >at the last dot (rawColumnName.lastIndexOf('.'))
> >By the way, is the fully qualified column name
> >SCHEMANAME.TABLENAME.COLUMNNAME in all Db's or are there exceptions ?
>
> And this is the reason why I need this. Even worse, I do need to
> switch the schema name on the fly.
>
> Do you already have code?
>
> Regards
> Henning
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
