BorlandTalk.com

Anders Balslev · Guest

Hi

BDS2006
WinXP
MSSQL 2005

Can anyone give me a hint?

I run an application connected to a MSSQL database.

Table1:
(Table1ID int IDENTITY(1,1),
Field2 ..,
Field3..,
MyDateTimeField datetime,
Field4..
..
.. )
Table1ID is the primary index (non-clustered)
MyDateTimeField is a secondary index (clustered)

Table2:
(Table2ID int IDENTITY(1,1),
Field2...,
Field3...,
Table1ID int)
Table2ID is the primary index (non-clustered)
Table1ID is the primary index (clustered)

Table1 and Table 2 are related in a Master-Detail relation

Both tables contains a "Lot of rows" (>100000)

In my DataModule I have
ADODataset1 (connected to Table1 by means of "SELECT * FROM Table1 WHERE
MyDateTimeField BETWEEN :DateFrom AND :DateTo"
- Opening this dataset shouldn't be a problem, limiting the dates

ADODataset2 (Connected to Table2 by means of "SELECT * FROM Table2") - and
here the problem comes.

I have ADODataset2-->MasterSource, pointing to a datasource, which points on
Table1, specifying the link field = Table1ID

These are agan concerted to ClientDatasets, where Clientdataset1 thereby
have a DatasetField

By analyzing the the SQL profiler-->Trace, I can see that table2 are opened
with ALL rows, not only the rows that are related to Table1.

Are there any parameters on the table components that I can set in order to
ask the compiler ONLY to return the detail rows that are relevant?

Best regards
Anders Balslev

Oliver Townshend · Guest