B
brynjar
Guest
Don't know if it's the best way and it's unfortunately not cost-based but we've put quite a few features into our query service. The programmer defines the query and optional sub-queries and/or rules for when to switch buffer-sequence. • Calculated fields (with optional runtime parameters) • Filter on calculated fields • Filter on joined buffers (subselect and not exists) • Ability to automatically change buffer sequence • Filter on external buffers (requires sub-query definition) • Multiple joins to same table • Sorting on calculated fields • Query count, count distinct • Accumulation of totals and subtotals • Support for duplicate field names • Support for extent-fields • Validation of query against schema security • Returns temp-table handle pluss additional context info (read/write restrictions) • Paging • Sub-selects (pre-scan) • All records (in query) are processed when • Sorting on calculated values • Querying distinct values • Grand totals involve calculated fields • All records are traversed when • Doing record count or grand totals
Continue reading...
Continue reading...