MetaDataManager init

rbuser

RB7.02 D6 Dade EndUser

I have a dictionary with quite a number of tables and columns and it
takes a few seconds for the MetaDataManager to initialize. I would
like to get that out of the way on program load - where the user is
used to waiting - instead of on first use (which causes an
unidentified pause for the user).

How can I do this? At the point where I want to do it I have the
dictionary etc. but I don't have a report loaded (and may have one at
all yet).

Thanks,
Rick Matthews
Dartek Systems Inc.

nicocizik

Rick,

There is a global MetaDataManager variable declared in the
daMetaDataManager. Try calling gMetaDataManager.GetTables to initialize the
MetaDataManager before its default initialization time. You will need to
create your own TList and TdaBDESession components to pass the GetTables
function. If you then need to retrieve the fields, you can loop through the
list you created and call GetFields on each of the tables.

--

rbuser

Thanks Nico, it was the need for a session component that stopped me
from do that. Is there anything special I need to do/set when I
create/destroy the component?

Thanks,

nicocizik

Rick,

Sorry for the delay. You should not have to do anything special when
creating an empty TdaBDESession object. I will work on an example that
should help you work through this. Thanks for your patience.

--

rbuser

Thanks Nico I have now done it.

However the reason I was doing this was to try to avoid a long pause
when the metadata was created. It seems though that my real problem
is the long time it takes to load each tables fields into the meta
data.

My setup is: RB7.02 D6 dade with ADO, dictionary. I have about 457
tables with fieldcounts ranging from 3 (lots) to some very wide
tables. Current average fields per table is 14.

I have a gMetaDataManager.GetTables followed by a loop through all of
the tables doing a gMetaDataManager.GetFields

Taking a random selection of tables it seems to take (on my Win2k box
under the debugger) about 60 seconds to do only 20 tables (ie 3
seconds per table).

Is this what your timings or performance goals would be? Am I doing
something wrong? How can I greatly decrease the time this takes?

Thanks,