查询真的很慢

Question

I am looking for some advice in relation to a query we run each week. The query takes far too long to return all rows (approx 16000) and takes roughly 1 hour. When you initially run the query, the 50 rows are brought back in a few mins, but if you press CTRL-A to fetch all rows, this is where it takes too much time. Is there anything in the query below i can amend to run faster? Thanks

SELECT
  Contract,
  Part_No,
  Description,
  Warehouse,
  Bay, 
  Location,
  HANDLING_UNIT_ID,
  Handling_Unit_API.Get_Alt_Handling_Unit_Label_Id(HANDLING_UNIT_ID) AS pallet_id,
  Lot_Batch_No,
  Qty,
  Uom,
  (Catch_Qty/1000) AS cATCH_QTY_T,
  Catch_Uom,
  Point_In_Time_Part_Cost AS Part_Cost,
  Qty * Point_In_Time_Part_Cost AS Value
  --Part_Value_Offset,
  --Qty * Point_In_Time_Part_Cost + Part_Value_Offset AS Expected_GL_Balance
FROM
(
SELECT
  s.Contract,
  s.Part_No,
  s.Description,
  s.Warehouse,
  s.Bay,
  s.Location,
  s.HANDLING_UNIT_ID,
  Handling_Unit_API.Get_Alt_Handling_Unit_Label_Id(s.HANDLING_UNIT_ID) AS ALT_HND,
  s.Lot_Batch_No,
  s.Qty,
  s.Uom,
  s.Catch_Qty,
  s.Catch_Uom,
  FM_Inventory_Util_Api.Get_PIT_Part_Cost(Contract, Part_No, Warehouse, To_Date('&EFFECTIVE_DATE', 'DD/MM/YYYY')) AS Point_In_Time_Part_Cost
FROM
(
SELECT
  ip.Contract,
  ip.Part_No,
  ip.Description,
  ip.Warehouse,
  ip.Bay_No AS Bay, 
  ip.Location_No AS Location,
  ip.HANDLING_UNIT_ID,
  Handling_Unit_API.Get_Alt_Handling_Unit_Label_Id(ip.HANDLING_UNIT_ID) AS ALT_HND,
  ip.Lot_Batch_No,
  SUM(ip.Qty) AS Qty,
  MIN(ip.UOM) AS UOM,
  SUM(ip.Catch_Qty) AS Catch_Qty,
  MIN(ip.Catch_UOM) AS Catch_UOM
FROM
(
SELECT
  ipis.Contract, 
  ipis.Part_No, 
  Inventory_Part_Api.Get_Description(ipis.Contract, ipis.Part_No) AS Description,
  ipis.Warehouse, 
  ipis.Bay_No, 
  ipis.Location_No,
  ipis.HANDLING_UNIT_ID,
  IFSAPP.Handling_Unit_API.Get_Alt_Handling_Unit_Label_Id(ipis.HANDLING_UNIT_ID) AS ALT_HND,
  ipis.Lot_Batch_No,
  SUM(ipis.Qty_Onhand) AS Qty,
  MIN(Inventory_Part_API.Get_Unit_Meas(ipis.Contract, ipis.Part_No)) UOM, 
  SUM(ipis.Catch_Qty_Onhand) AS Catch_Qty,
  (SELECT unit_code
      FROM   PART_CATALOG_TAB
      WHERE  part_no = IPIS.part_no
      AND cATCH_UNIT_ENABLED = 'TRUE') AS CATCH_UOM
FROM
  Inventory_Part_In_Stock ipis
WHERE
  ipis.Contract LIKE Nvl('&SITE', '%')
AND  ipis.Warehouse LIKE Nvl('&WAREHOUSE', '%')
AND  (ipis.Bay_No LIKE Nvl('&BAY', '%') or ipis.Bay_No is null)

GROUP BY
  ipis.Contract, 
  ipis.Part_No, 
  Inventory_Part_Api.Get_Description(ipis.Contract, ipis.Part_No),
  ipis.Warehouse, 
  ipis.Bay_No, 
  ipis.Location_No,
  ipis.HANDLING_UNIT_ID,
  ifsapp.Handling_Unit_API.Get_Alt_Handling_Unit_Label_Id(ipis.HANDLING_UNIT_ID),
  ipis.Lot_Batch_No
HAVING 
  SUM(ipis.Qty_Onhand)  <> 0
UNION ALL
SELECT DISTINCT
  i.Contract, 
  i.Part_No,
  Inventory_Part_Api.Get_Description(i.Contract, i.Part_No) AS Description,
  Inventory_Location_API.Get_Warehouse(i.Contract, i.Location_No) AS Warehouse,
  Inventory_Location_API.Get_Bay_No(i.contract, i.location_no) AS Bay_No,
  i.Location_No,
  i.HANDLING_UNIT_ID,
  Handling_Unit_API.Get_Alt_Handling_Unit_Label_Id(i.HANDLING_UNIT_ID) AS ALT_HND,
  i.Lot_Batch_No,
  Nvl(SUM(Decode(i.Direction, '+', 0 - i.Quantity, i.Quantity)), 0) AS Qty,
  MIN(Inventory_Part_API.Get_Unit_Meas(i.Contract, i.Part_No)) UOM,
  Nvl(SUM(Decode(i.Catch_Direction, '+', 0 - i.Catch_Quantity, i.Catch_Quantity)), 0) AS Catch_Qty,
  (SELECT unit_code
      FROM   PART_CATALOG_TAB
      WHERE  part_no = i.part_no
      AND cATCH_UNIT_ENABLED = 'TRUE') AS CATCH_UOM
FROM
  Inventory_Transaction_Hist2 i
WHERE
  i.Contract LIKE Nvl('&SITE', '%')
AND
  i.Date_Applied BETWEEN To_Date('&EFFECTIVE_DATE', 'DD/MM/YYYY') + 1 AND Trunc(SYSDATE)
AND
  i.Direction IN ('+', '-')
AND
  Inventory_Location_API.Get_Warehouse(i.Contract, i.Location_No) LIKE Nvl('&WAREHOUSE', '%')
AND
  (Inventory_Location_API.Get_Bay_No(i.contract, i.location_no) LIKE Nvl('&BAY', '%') or 
   inventory_Location_API.Get_Bay_No(i.contract, i.location_no) is null)
GROUP BY
  i.Contract, 
  i.Part_No,
  Inventory_Part_Api.Get_Description(i.Contract, i.Part_No),
  Inventory_Location_API.Get_Warehouse(i.Contract, i.Location_No),
  Inventory_Location_API.Get_Bay_No(i.contract, i.location_no),
  i.Location_No,
  i.HANDLING_UNIT_ID,
  Handling_Unit_API.Get_Alt_Handling_Unit_Label_Id(i.HANDLING_UNIT_ID),
  i.Lot_Batch_No
HAVING 
  Nvl(SUM(Decode(i.Direction, '+', 0 - i.Quantity, i.Quantity)), 0)  <> 0
) ip
GROUP BY
  ip.Contract,
  ip.Part_No,
  ip.Description,
  ip.Warehouse,
  ip.Bay_No,
  ip.Location_No,
  ip.HANDLING_UNIT_ID,
  Handling_Unit_API.Get_Alt_Handling_Unit_Label_Id(ip.HANDLING_UNIT_ID),
  ip.Lot_Batch_No
HAVING
  SUM(ip.Qty) <> 0
) s
)

Answer 1

Based on my experience, subqueries could be deadly performance-wise, especially if you are dealing with a large amount of data. If the problem could not be handled by joining tables, I would advise pre-defining the subqueries within a WITH statement and then referring to those instead. It will significantly increase your querying speed, trust me. You can define multiple queries inside one WITH statement. More details on using the WITH statement here