Need help query for current row and previous row

  • Hye,

    I've table and data as following,

    CREATE TABLE [dbo].[WeightStory](
    [Idx] [int] IDENTITY(1,1) NOT NULL,
    [WeightDte] [datetime] NULL,
    [WhatWeight] [decimal](18, 3) NULL,
    CONSTRAINT [PK_WeightStory] PRIMARY KEY CLUSTERED
    (
    [Idx] ASC
    )WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
    ) ON [PRIMARY]

    GO
    SET IDENTITY_INSERT [dbo].[WeightStory] ON

    GO
    INSERT [dbo].[WeightStory] ([Idx], [WeightDte], [WhatWeight]) VALUES (1, CAST(N'2019-10-12 11:00:00.000' AS DateTime), CAST(99.200 AS Decimal(18, 3)))
    GO
    INSERT [dbo].[WeightStory] ([Idx], [WeightDte], [WhatWeight]) VALUES (2, CAST(N'2019-10-14 16:00:00.000' AS DateTime), CAST(99.500 AS Decimal(18, 3)))
    GO
    INSERT [dbo].[WeightStory] ([Idx], [WeightDte], [WhatWeight]) VALUES (3, CAST(N'2019-10-15 07:58:00.000' AS DateTime), CAST(98.400 AS Decimal(18, 3)))
    GO
    INSERT [dbo].[WeightStory] ([Idx], [WeightDte], [WhatWeight]) VALUES (4, CAST(N'2019-10-16 06:57:00.000' AS DateTime), CAST(98.600 AS Decimal(18, 3)))
    GO
    INSERT [dbo].[WeightStory] ([Idx], [WeightDte], [WhatWeight]) VALUES (5, CAST(N'2019-10-17 08:00:00.000' AS DateTime), CAST(97.900 AS Decimal(18, 3)))
    GO
    SET IDENTITY_INSERT [dbo].[WeightStory] OFF
    GO

    My query will return as follow,

    select [WeightDte], [WhatWeight] from WeightStory order by WeightDte desc

    WeightDteWhatWeight
    -------------------------------------
    2019-10-17 08:00:00.00097.900
    2019-10-16 06:57:00.00098.600
    2019-10-15 07:58:00.00098.400
    2019-10-14 16:00:00.00099.500
    2019-10-12 11:00:00.00099.200

    How to query to get result as following,

    WeightDteWhatWeight                   Accomplishment
    ---------------------------------------------------------------
    2019-10-17 08:00:00.00097.900 Down
    2019-10-16 06:57:00.00098.600 Up
    2019-10-15 07:58:00.00098.400 Down
    2019-10-14 16:00:00.00099.500 Up
    2019-10-12 11:00:00.00099.200 N/A

    You will see, to display Accomplishment Up or Down - The current row(WhatWeight) will depend on previous row(WhatWeight)

    Please help

  • Use the LAG or LEAD function.  Try it yourself, and post back if there's anything in particular that you struggle with.

    John

  • Hello John,

    SELECT 
    [WeightDte],
    [WhatWeight],
    LEAD([WhatWeight],1) OVER (
    PARTITION BY WhatWeight
    ORDER BY [WeightDte] desc
    ) LastWeight
    FROM
    [dbo].[WeightStory]
    Order by [WeightDte] desc


    WeightDteWhatWeightLastWeight
    --------------------------------------------------------
    2019-10-17 08:00:00.00097.900NULL
    2019-10-16 06:57:00.00098.600NULL
    2019-10-15 07:58:00.00098.400NULL
    2019-10-14 16:00:00.00099.500NULL
    2019-10-12 11:00:00.00099.200NULL

    Please help

  • Remove this part

    PARTITION BY WhatWeight
  • SELECT 
    [WeightDte],
    [WhatWeight],
    LEAD([WhatWeight],1) OVER (

    ORDER BY [WeightDte] desc
    ) LastWeight
    FROM
    [dbo].[WeightStory]
    Order by [WeightDte] desc


    WeightDteWhatWeightLastWeight
    -----------------------------------------------------
    2019-10-17 08:00:00.00097.90098.600
    2019-10-16 06:57:00.00098.60098.400
    2019-10-15 07:58:00.00098.40099.500
    2019-10-14 16:00:00.00099.50099.200
    2019-10-12 11:00:00.00099.200NULL

    Yippieeeeeeeeee. Thanks John and DesNorton

    • This reply was modified 4 years, 6 months ago by  Adelia.
  • I went on the assumption you wanted the UP & DOWN.

    SELECT
    [WeightDte],
    [WhatWeight],
    CASE
    WHEN LEAD([WhatWeight],1) OVER (ORDER BY [WeightDte] DESC) > WhatWeight THEN 'DOWN'
    WHEN LAG([WhatWeight],1) OVER (ORDER BY [WeightDte] DESC) < WhatWeight THEN 'UP'
    ELSE 'N/A' END AS LastWeight
    FROM [dbo].[WeightStory]Order by [WeightDte] desc

    • This reply was modified 4 years, 6 months ago by  inkosi_dba_dev.
    • This reply was modified 4 years, 6 months ago by  inkosi_dba_dev.
    • This reply was modified 4 years, 6 months ago by  inkosi_dba_dev. Reason: Missed a ")" on second line of case statement
  • In case you're not aware LEAD(WhatWeight,1) OVER (ORDER BY WeightDte desc) is equivalent to LAG(WhatWeight,1) OVER (ORDER BY WeightDte).  In this case, it is right to use LEAD() because the sort order matches the sort order of the overall query, so you only need one sort operation.  Otherwise, I would recommend using ascending order.

    Drew

    J. Drew Allen
    Business Intelligence Analyst
    Philadelphia, PA

  • Drew,

    To better understand can you please supply a re-structured query that would achieve the same results. This would help to put a method to result. Which I assume is just replace the LAG portion in the second CASE line with the LEAD equivalent. Can you post the Execution Plan for both.

  • inkosi_dba_dev wrote:

    Drew,

    To better understand can you please supply a re-structured query that would achieve the same results. This would help to put a method to result

    You learn more by trying it yourself.  You have all the information that you need.  I'm not going to spoon feed you something that you should be able to figure out on your own.

    Drew

    J. Drew Allen
    Business Intelligence Analyst
    Philadelphia, PA

  • Drew,

    No one is asking you to spoon feed anything, you made a comment so back it up with the intend results. If you wanted to contribute to the question by trying to pick at something then say so. I already did the work and posted my intend results.

Viewing 10 posts - 1 through 9 (of 9 total)

You must be logged in to reply to this topic. Login to reply