Fast Query

[Kivju]

Is possible make faster query ?
I have follow code and this is slow.
Different system ?


CString Day= "20040524";
CString Period1 ="Other1";
CString Period2 ="Other2";

m_pSet->Close();
m_pSet->m_strFilter.Format(" COL1 > '%s' AND COL2 <> '%s' AND COL3 <> '%s'" ,Day,Period1 ,Period2);
m_pSet->Open();

int recCount = m_pSet->GetRecordCount();
if (recCount == 0){
m_pSet->Close();
m_pSet->m_strFilter = "";
m_pSet->Open();
nFields = 0;
m_list.DeleteAllItems();}
else { int ii = 0;
}

while ( !m_pSet->IsEOF( ) ){

m_pSet->GetFieldValue(("COL1"), CDay);
m_list.InsertItem(ii, CDay);

m_pSet->GetFieldValue(("COL2"), CPeriod1);
m_list.SetItemText(ii, 1, CPeriod1);

m_pSet->GetFieldValue(("COL3"), CPeriod2);
m_list.SetItemText(ii, 2, CPeriod2);

m_pSet->GetFieldValue(("COL4"), CPeriod3);
m_list.SetItemText(ii, 3, CPeriod3);

UpdateData(FALSE);
m_pSet->MoveNext();}

[TheCPUWizard]

If you want any type of performance, do NOT use alpha column lookups within the loop....

[Kivju]

i'm stupid, could you elaborate ?

[GCDEF]

Documentation for GetRecordCount says you have to actually read the records. Just calling it after an open won't work. Test for IsEOF() instead. Other than that, just make sure the fields you're querying are indexed. Also, if it's a big table, a dynaset may run quicker than a snapshot.