Some body please help about the Inventory FIFO (FIRST IN FISRT OUT ) query is very slow

Question

I have table ''tran_stock''  to store a transaction (IN-OUT) of the stock. 
The problem is when the  transaction rows are more than 50000 the query to get the balance
stock (Total 'IN' - Total'Out' by FIFO) is very slow . Somebody please suggest me to do this.
This is my query
;WITH OrderedIn as (
    select *,ROW_NUMBER() OVER (PARTITION BY Stock_ID ORDER BY TranDate) as rn
    from Tran_Stock
    where TxnType = 'IN'  
), RunningTotals as (
    select  Stock_ID ,Qty,Price,Qty as Total,Cast(0 as decimal(10,2)) as PrevTotal ,  trandate , rn  from OrderedIn where rn = 1
    union all
    select  rt.Stock_ID,oi.Qty,oi.Price,Cast(rt.Total + oi.Qty as decimal(10,2)),Cast(rt.Total as decimal(10,2)) , oi.TranDate ,oi.rn 
    from
        RunningTotals rt
            inner join
        OrderedIn oi
            on
                rt.Stock_ID = oi.Stock_ID and
                rt.rn = oi.rn - 1
, TotalOut as (
select Stock_ID ,SUM(Qty) as Qty from Tran_Stock where TxnType='OUT'   group by Stock_ID
) ,  GrandTotal as 
select
     rt.Stock_ID , rt.QTY AS original ,SUM(CASE WHEN PrevTotal > isNULL(out.Qty, 0) THEN rt.Qty ELSE rt.Total - isNULL(out.Qty, 0) END) AS QTY , Price , SUM(CASE WHEN PrevTotal > isNULL(out.Qty, 0) THEN rt.Qty ELSE rt.Total - isNULL(out.Qty,
0) END * Price) AS Ending , rt.TranDate
from
    RunningTotals rt
        LEFT join
    TotalOut out
        on
            rt.Stock_ID = out.Stock_ID
where rt.Total > isNull(out.Qty , 0)
group by rt.Stock_ID , Price , rt.TranDate  , rt.QTY
) SELECT * FROM  GrandTotal  order by TranDate  option (maxrecursion 0)
AND  this is my Table with some example data
USE [TestInventory]
GO
/****** Object:  Table [dbo].[Tran_Stock]    Script Date: 12/15/2014 3:55:56 AM ******/
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
CREATE TABLE [dbo].[Tran_Stock](
[TranID] [int] IDENTITY(1,1) NOT NULL,
[Stock_ID] [int] NULL,
[TranDate] [date] NULL,
[TxnType] [nvarchar](3) NULL,
[Qty] [decimal](10, 2) NULL,
[Price] [decimal](10, 2) NULL
) ON [PRIMARY]
GO
SET IDENTITY_INSERT [dbo].[Tran_Stock] ON 
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (1, 1, CAST(0x81350B00 AS Date), N'IN', CAST(200.00 AS Decimal(10, 2)), CAST(750.00 AS Decimal(10, 2)))
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (2, 1, CAST(0x85350B00 AS Date), N'OUT', CAST(100.00 AS Decimal(10, 2)), NULL)
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (3, 1, CAST(0x8A350B00 AS Date), N'IN', CAST(50.00 AS Decimal(10, 2)), CAST(700.00 AS Decimal(10, 2)))
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (4, 1, CAST(0x90350B00 AS Date), N'IN', CAST(75.00 AS Decimal(10, 2)), CAST(800.00 AS Decimal(10, 2)))
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (5, 1, CAST(0x99350B00 AS Date), N'OUT', CAST(175.00 AS Decimal(10, 2)), NULL)
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (6, 2, CAST(0x82350B00 AS Date), N'IN', CAST(150.00 AS Decimal(10, 2)), CAST(350.00 AS Decimal(10, 2)))
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (7, 2, CAST(0x88350B00 AS Date), N'OUT', CAST(40.00 AS Decimal(10, 2)), NULL)
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (8, 2, CAST(0x8C350B00 AS Date), N'OUT', CAST(10.00 AS Decimal(10, 2)), NULL)
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (9, 2, CAST(0x98350B00 AS Date), N'IN', CAST(90.00 AS Decimal(10, 2)), CAST(340.00 AS Decimal(10, 2)))
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (12, 2, CAST(0x98350B00 AS Date), N'IN', CAST(10.00 AS Decimal(10, 2)), CAST(341.00 AS Decimal(10, 2)))
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (13, 2, CAST(0x99350B00 AS Date), N'OUT', CAST(30.00 AS Decimal(10, 2)), NULL)
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (14, 2, CAST(0x81350B00 AS Date), N'IN', CAST(120.00 AS Decimal(10, 2)), CAST(350.00 AS Decimal(10, 2)))
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (15, 2, CAST(0x89350B00 AS Date), N'OUT', CAST(90.00 AS Decimal(10, 2)), NULL)
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (17, 3, CAST(0x89350B00 AS Date), N'IN', CAST(90.00 AS Decimal(10, 2)), CAST(350.00 AS Decimal(10, 2)))
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (18, 3, CAST(0x8A350B00 AS Date), N'OUT', CAST(60.00 AS Decimal(10, 2)), NULL)
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (19, 3, CAST(0x5D390B00 AS Date), N'OUT', CAST(20.00 AS Decimal(10, 2)), NULL)
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (20, 1, CAST(0x81350B00 AS Date), N'IN', CAST(200.00 AS Decimal(10, 2)), CAST(750.00 AS Decimal(10, 2)))
INSERT [dbo].[Tran_Stock] ([TranID], [Stock_ID], [TranDate], [TxnType], [Qty], [Price]) VALUES (21, 1, CAST(0x85350B00 AS Date), N'OUT', CAST(100.00 AS Decimal(10, 2)), NULL)

Answer 1

Hi,
You dont have any Index on the table, so the filters (where part in the queries) and the sorting (order by, group by) make your query slow. it take time to sort or filter data without Index. It is like reading a book with 3k pages and trying to find a specific
subject. You have to read the all books, but if you have "table of contents" which is the book index, then this can be done fast.
* check the Execution plan (the image here show part of your query's execution plan). the query scan the all table 3 times...
there is one sort that use 81% of the query resources... I will go sleep soon (it is
23:58 In Israel now), and I don't know if I will have time to read the query itself and improved it, so lets start with indexes :-) 
** as a first step after or before you create the right Indexes you should remove the recursive all together. there is no reason for lopping the data when you can use it as a SET. This is where the relational database's power come to life :-)
for example check this code. I only use your first part of the code using the first 2 CTE tables. compare result.
;WITH
OrderedIn as (
select *,ROW_NUMBER() OVER (PARTITION BY Stock_ID ORDER BY TranDate) as rn
from Tran_Stock
where TxnType = 'IN'
, RunningTotals as (
select Stock_ID ,Qty, Price, Qty as Total, Cast(0 as decimal(10,2)) as PrevTotal, trandate , rn
from OrderedIn where rn = 1
union all
select
rt.Stock_ID, oi.Qty, oi.Price, Cast(rt.Total + oi.Qty as decimal(10,2))
,Cast(rt.Total as decimal(10,2))
, oi.TranDate ,oi.rn
from RunningTotals rt
inner join OrderedIn oi on rt.Stock_ID = oi.Stock_ID and rt.rn = oi.rn - 1
select * from RunningTotals
-- This will do the same as the above query
select
Stock_ID, Qty, Price
, SUM(Qty) over (PARTITION BY Stock_ID order by TranDate ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) as Total
,TranDate--,TranID, TxnType
from Tran_Stock
where TxnType = 'IN'
order by Stock_ID
  Ronen Ariely
 [Personal Site]    [Blog]    [Facebook]