SQL Merge into getting stuck

Question

I want to Merge into some data from a (python) panda dataframe to my sql-server 2019 running on azure.

This is my code with example data values:

MERGE INTO [dbo].[APPSCHED_CNG] AS t
USING (
    VALUES
        (53243, '2023-06-19 12:39:35', 20230621, 80000, 20230621, 90000, 0, 0),(53259, '2023-06-19 12:39:35', 20230621, 123000, 20230621, 133000, 0, 0),(53266, '2023-06-19 12:39:35', 20230621, 93000, 20230621, 103000, 0, 0),(53234, '2023-06-19 12:39:35', 20230621, 80000, 20230621, 80000, 0, 0),(53286, '2023-06-19 12:39:35', 20230621, 80000, 20230621, 120000, 0, 0),(53272, '2023-06-19 12:39:35', 20230621, 123000, 20230621, 153000, 0, 0),(51509, '2023-06-19 12:39:35', 20230621, 74500, 20230621, 74500, 0, 0),(53247, '2023-06-19 12:39:35', 20230621, 80000, 20230621, 110000, 0, 0),(53255, '2023-06-19 12:39:35', 20230621, 80000, 20230621, 80000, 0, 0),(53236, '2023-06-19 12:39:35', 20230621, 93000, 20230621, 123000, 0, 0),(53251, '2023-06-19 12:39:35', 20230621, 93000, 20230621, 113000, 0, 0),(52915, '2023-06-19 12:39:35', 20230621, 100000, 20230621, 110000, 0, 0),(53238, '2023-06-19 12:39:35', 20230621, 100000, 20230621, 120000, 0, 0),(53239, '2023-06-19 12:39:35', 20230621, 123000, 20230621, 143000, 0, 0),(53284, '2023-06-19 12:39:35', 20230621, 143000, 20230621, 153000, 0, 0)
) AS s ([KEY], [change], [ANNDAT], [ANNZEIT], [ENDDAT], [ENDZEIT], [RG], [DATAGRP])
ON t.[KEY] = s.[KEY] AND t.[ANNDAT] = s.[ANNDAT] AND t.[ANNZEIT] = s.[ANNZEIT] AND t.[ENDDAT] = s.[ENDDAT] AND t.[ENDZEIT] = s.[ENDZEIT] AND t.[DATAGRP] = s.[DATAGRP]
WHEN NOT MATCHED THEN
    INSERT ([KEY], [change], [ANNDAT], [ANNZEIT], [ENDDAT], [ENDZEIT], [RG], [DATAGRP])
    VALUES (s.[KEY], s.[change], s.[ANNDAT], s.[ANNZEIT], s.[ENDDAT], s.[ENDZEIT], s.[RG], s.[DATAGRP]);

What it should do: I have this data from a IBM db2 database. When I open the frontend, the backend should check DB2 for new datasets. If the datasets are not on the Azure-SQL, the new datasets should be inserted - exept, if there is a difference only at RG, so a different RG is not relevant if there is a difference.

And here is the table script:

SET ANSI_NULLS ON
GO

SET QUOTED_IDENTIFIER ON
GO

CREATE TABLE [dbo].[APPSCHED_CNG](
    [KEY] [int] NULL,
    [change] [datetime] NULL,
    [ANNDAT] [int] NULL,
    [ANNZEIT] [int] NULL,
    [ENDDAT] [int] NULL,
    [ENDZEIT] [int] NULL,
    [RG] [int] NULL,
    [DATAGRP] [bit] NULL
) ON [PRIMARY]
GO

Sometimes it works, sometimes it causes the SSMS to load and load and load and it looks it do not find an end..

What seems to be the problem with my code? Is it a problem that I did not support WHEN MATCHED THEN?

Thank you

Answer 1

INSERT INTO SELECT WHERE NOT EXISTS as written in the comments brings the solution. Here updated code:

INSERT INTO [dbo].[APPSCHED_CNG] ([KEY], [change], [ANNDAT], [ANNZEIT], [ENDDAT], [ENDZEIT], [RG], [DATAGRP])
                        SELECT s.[KEY], s.[change], s.[ANNDAT], s.[ANNZEIT], s.[ENDDAT], s.[ENDZEIT], s.[RG], s.[DATAGRP]
                        FROM (
                            VALUES
                                (51509, '2023-06-19 15:59:31', 20230621, 80000, 20230621, 74500, 0, 0),
                                (53247, '2023-06-19 15:57:31', 20230621, 80000, 20230621, 110000, 0, 0)
                                --rest of values            
                        ) AS s ([KEY], [change], [ANNDAT], [ANNZEIT], [ENDDAT], [ENDZEIT], [RG], [DATAGRP])
                        WHERE NOT EXISTS (
                            SELECT 1
                            FROM [dbo].[APPSCHED_CNG] t
                            WHERE t.[KEY] = s.[KEY]
                                AND t.[ANNDAT] = s.[ANNDAT]
                                AND t.[ANNZEIT] = s.[ANNZEIT]
                                AND t.[ENDDAT] = s.[ENDDAT]
                                AND t.[ENDZEIT] = s.[ENDZEIT]
                                AND t.[DATAGRP] = s.[DATAGRP]
                        );