I want to know how simply can write the code below and also i am interested to avoid code repetition. Is there any other way. Do's are
Whatever changes in the fields of [invoices_global] table, that changes would be inserted as new records in the format of old_value, new_value, field_changed in the Change log table.
A single field may be updated by many users, so latest should take, thats why i mentioned as last_updated_datetime desc in the code.
Need to avoid repetition.
Performance improvement, since i am calling the same table again and again multiple times for single ap_work_id. It takes 18sec to fetch results - 2000 out of 40000.
FYI, system_id and ap_work_id are same.
SELECT api.ap_work_id, api.invoice_status, api.po_number, api.po_value, api.region, api.invoice_value, api.invoice_value_currency, em3.employee_full_name AS invoice_owner, ( **SELECT TOP 1 old_value FROM change_log cl WHERE field_changed = 'ARIBA INVOICE ID' AND api.ap_work_id = cl. system_id AND system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS bot_ariba_id_value, ( **SELECT TOP 1 new_value FROM change_log cl WHERE field_changed = 'ARIBA INVOICE ID' AND api.ap_work_id = cl. system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS processor_ariba_id_value, ( **SELECT TOP 1 old_value FROM change_log cl WHERE field_changed = 'PO NUMBER' AND api.ap_work_id = cl.system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS bot_po_number, ( **SELECT TOP 1 new_value FROM change_log cl WHERE field_changed = 'PO NUMBER' AND api.ap_work_id = cl.system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS processor_po_number, ( **SELECT TOP 1 old_value FROM change_log cl WHERE field_changed = 'VENDOR NAME' AND api.ap_work_id = cl.system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS bot_vendor_name, ( **SELECT TOP 1 new_value FROM change_log cl WHERE field_changed = 'VENDOR NAME' AND api.ap_work_id = cl.system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS processor_vendor_name, ( **SELECT TOP 1 old_value FROM change_log cl WHERE field_changed = 'INVOICE VALUE' AND api.ap_work_id = cl.system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS bot_invoice_value, ( **SELECT TOP 1 new_value FROM change_log cl WHERE field_changed = 'INVOICE VALUE' AND api.ap_work_id = cl.system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS processor_invoice_value, ( **SELECT TOP 1 old_value FROM change_log cl WHERE field_changed = 'INVOICE CURRENCY' AND api.ap_work_id = cl. system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS bot_invoice_currency, ( **SELECT TOP 1 new_value FROM change_log cl WHERE field_changed = 'INVOICE CURRENCY' AND api.ap_work_id = cl. system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS processor_invoice_currency, ( **SELECT TOP 1 old_value FROM change_log cl WHERE field_changed = 'INVOICE DATE' AND api.ap_work_id = cl.system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS bot_invoice_date, ( **SELECT TOP 1 new_value FROM change_log cl WHERE field_changed = 'INVOICE DATE' AND api.ap_work_id = cl.system_id AND cl.system_type = 'INVOICE' ORDER BY last_updated_datetime DESC** ) AS processor_invoice_date FROM invoices_global api OUTER APPLY ( SELECT TOP 1 * FROM transactions t WHERE t.po_number = api.po_number AND t.po_number <> '' ORDER BY t.last_updated_datetime DESC ) AS t LEFT JOIN employee_master em3 ON em3.ldap = api.invoice_owner_ldap LEFT JOIN vendor v ON v.vendor_number = api.vendor_number LEFT JOIN cost_center cc ON cc.cost_ctr = api.cost_center WHERE DATEDIFF(DAY,api.approved_date,GETDATE()) <= 30 GROUP BY api.ap_work_id, api.category, api.invoice_status, api.po_number, api.po_value, api.region, em3.employee_full_name, api.invoice_value, api.invoice_value_currency, api.invoice_date, api.due_date, api.company_code, api.vendor_number, api.approved_date, t.company_code, api.cost_center, v.vendor_name, v.vendor_number, cc.company;
Here's an approach where you
It currently uses MAX(CASE()) but you could do this as a PIVOT.
Only the top part of the SQL (relevant to the problem) is included here - you'll need to do the other joins on the bottom. I also haven't tested it (without table structures etc) so I can't promise there are no typos.