Introduction
Systems often work with distributed databases that contain distributed tables. Distribution is supported by various mechanisms, including replication. In this case, it is necessary to constantly maintain the synchronization of a particular data segment. It is also necessary to check the synchronization itself. This is when the need to compare data in two tables appears.
Before comparing data in two tables, you need to make sure that the schemas of the compared tables are either the same or acceptably different. By acceptably different, we mean a difference in the definition of two tables, whereby data can be compared correctly. For example, types of corresponding columns of compared tables should be mapped without data loss in these columns.
Let's compare the SQL Server schemas of the two Employee tables from two different databases JobEmpl and JobEmplDB.
For further work, it is necessary to recall the definitions of the Employee table of JobEmpl and JobEmplDB databases:
USE [JobEmpl]
GO
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
CREATE TABLE [dbo].[Employee](
[EmployeeID] [int] IDENTITY(1,1) NOT NULL,
[FirstName] [nvarchar](255) NOT NULL,
[LastName] [nvarchar](255) NOT NULL,
[Address] [nvarchar](max) NULL,
[CheckSumVal] AS (checksum((coalesce(CONVERT([nvarchar](max),[FirstName]),N'')+coalesce(CONVERT([nvarchar](max),[LastName]),N''))+coalesce(CONVERT([nvarchar](max),[Address]),N''))),
[REPL_GUID] [uniqueidentifier] ROWGUIDCOL NOT NULL,
CONSTRAINT [PK_Employee_EmployeeID] PRIMARY KEY CLUSTERED
(
[EmployeeID] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON, FILLFACTOR = 80) ON [PRIMARY]
) ON [PRIMARY] TEXTIMAGE_ON [PRIMARY]
GO
ALTER TABLE [dbo].[Employee] ADD CONSTRAINT [Employee_DEF_REPL_GUID] DEFAULT (newsequentialid()) FOR [REPL_GUID]
GO
//and
USE [JobEmplDB]
GO
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
CREATE TABLE [dbo].[Employee](
[EmployeeID] [int] IDENTITY(1,1) NOT NULL,
[FirstName] [nvarchar](255) NOT NULL,
[LastName] [nvarchar](255) NOT NULL,
[Address] [nvarchar](max) NULL,
CONSTRAINT [PK_Employee_EmployeeID] PRIMARY KEY CLUSTERED
(
[EmployeeID] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY] TEXTIMAGE_ON [PRIMARY]
GO
Comparing Database Schemas using SQL Server Data Tools
With the help of Visual Studio and SSDT, you can compare database schemas. To do this, you need to create a new project "JobEmployee" by doing the following:
Img.1. Creating a project JobEmployeeDB
Then you need to import the database.
To do this, right-click the project and in the context menu, select Import \ Database..:
Img.2. Select database import
Next, hit the "Select connection…" button and in the cascading menu, in the "Browse" tab set up the connection to JobEmpl database as follows:
Img.3. Configuring JobEmpl Database Import
Next, click the "Start" button to start the import of the JobEmpl database:
Img.4. Starting the database import process
You will then see a window showing the progress of the database import:
Img.5. The window of the database import progress
When the database import process is completed, press "Finish":
Img.6. Finishing database import
Once it is finished, JobEmployee project will contain directories, subdirectories, and database objects definitions in the following form:
Img.7.JobEmployee project after importing the JobEmplDB database
In the same way, we create a similar JobEmployeeDB project and import JobEmplDB database into it:
Img.8. JobEmployeeDB project after JobEmplDB database import
Now, right-click the JobEmployee project and in the drop-down menu, select "Schema Compare":
Img.9. Calling the database schema compare window
This will bring up the database schema compare window.
In the window, you need to select the projects as source and target, and then click the "Compare" button to start the comparison process:
Img.10. Database schema comparison window
We can see here that despite the differences between the definitions of the Employee tables in two databases, the table columns that we need for comparison are identical in data type. This means that the difference in the schemas of the Employee tables is acceptable. That is, we can compare the data in these two tables.
We can also use other tools to compare database schemas such as dbForge Schema Compare for SQL Server.
Comparing database schemas with the help of dbForge Schema Compare
Now, to compare database table schemas, we use a tool dbForge Schema Compare for SQL Server, which is also included in SQL Tools.
For this, in SSMS, right-click the first database and in the drop-down menu, select Schema Compare\ Set as Source:
Img.11. Selecting the source-base for schema comparison
We simply transfer JobEmplDB, the second database, to Target area and click the green arrow between source and target:
Img.12. Setting up database schema comparison
You simply need to press the "Next" button in the opened database schema comparison project:
Img.13. Choosing the source and target for schema comparison
Leave the following settings at their defaults and click the "Next" button:
Img.14. Schema comparison options
In the "Schema Mapping" tab, we also leave everything by default and press the "Next" button:
Img.15. "Schema Mapping" Tab
On the "Table Mapping" tab, select the required Employee table and on the right of the table name, click the ellipsis:
Img.16. Selecting the Employee table
The table mapping window opens up:
Img.17. Column mapping of two tables
In our case, only 4 fields are mapped, because two last fields are contained only in the JobEmpl database and are absent in the JobEmplDB database.
This setting is useful when column names in the source table and target table do not match.
The "Column details" table displays the column definition details in two tables: on the left – from the source database and on the right – from the target database.
Now hit the "OK" button
Img.18. Column mapping of two tables-2
Now, to start the database schema comparison process, click the "Compare" button:
Img.19. Starting schema comparison
A progress bar will appear:
Img.20. Schema comparison progress
We then select the desired Employee table,
Img.21. The comparison of Employee table schema
At the bottom left, you can see the code for defining the source database table and on the right – the target database table.
We can see here, as before, that the definitions of the Employee table in two databases JobEmpl and JobEmplDB show admissible distinction, that is why we can compare data in these two tables.
Let us now move on to the comparison of the data in two tables itself.
Comparing database data using SSIS
Let’s first make a comparison using SSIS. For this, you need to have SSDT installed.
We create a project called Integration Service Project in Visual Studio and name it IntegrationServicesProject
Img.22. Creating a project called Integration Services Project
We then create three connections:
- To the source JobEmpl database
- To the target JobEmplDB database
- To the JobEmplDiff database, where the table of differences will be displayed the following way below:
Img.23. Making a database connection
That way, new connections will be displayed in the project, as follows:
Img.24. Displaying the created connections
Then, in the project, in the "Control Flow" tab, we create a data flow task and name it "data flow task",
Img.25. Creating a data flow task
Let us now switch to the data flow and create an element "Source OLE DB" by doing the following:
Img.26. Creating a data source
On the "Columns" tab, we then select the fields required for comparison,
Img.27. Selecting the required fields
Next, we leave by default and press the "OK" button,
Img.28."Error Output" Tab
And now, right-click the created data source and in the drop-down menu, select "Show Advanced Editor…"
Img.29. Selecting "Show Advanced Editor…"
On the "Input and Output Properties" tab, set IsSorted to True in the properties for OLE DB source output and OLE DB source errors output,
Img.30. Setting IsSorted property to True
Next, for each of the "Output Columns" groups for the EpmloyeeID column, set SortKeyPosition property to 1. That is, we sort by the EmployeeID field value in ascending order,
Img.31. Choosing sort by the EmployeeID field value in ascending order
Similarly, let us create and set the data source to the JobEmplDB database.
That way, we obtain two created sources in the data flow task:
Img.32. Created data sources
Now, we create a merge join element in the following way:
Img.33. Creating "Merge Join"
Please note that we merge tables using a full outer join.
We then connect our sources to the created join element by merging "Merge Join", as follows:
Img.34. Connecting sources to "Merge Join"
We make the connection from JobEmpl left and the connection from JobEmplDB – right.
In fact, it is not that important, it is possible to do this the other way around.
In the JobEmplDiff database, we create a different table called EmployeeDiff, where we are going to put data differences in the following manner:
USE [JobEmplDiff]
GO
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
CREATE TABLE [dbo].[EmployeeDiff](
[ID] [int] IDENTITY(1,1) NOT NULL,
[EmployeeID] [int] NULL,
[EmployeeID_2] [int] NULL,
[FirstName] [nvarchar](255) NULL,
[FirstName_2] [nvarchar](255) NULL,
[LastName] [nvarchar](255) NULL,
[LastName_2] [nvarchar](255) NULL,
[Address] [nvarchar](max) NULL,
[Address_2] [nvarchar](max) NULL,
CONSTRAINT [PK_EmployeeDiff_1] PRIMARY KEY CLUSTERED
(
[ID] ASC
) WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON, FILLFACTOR = 80) ON [PRIMARY]
) ON [PRIMARY] TEXTIMAGE_ON [PRIMARY]
GO
Now, let us get back to our project and in the data flow task, we create a conditional split element as follows:
Img.35. Creating a conditional split
In the Conditional field for NotMatch, you need to type the following expression:
(
ISNULL(EmployeeID)
|| ISNULL(EmployeeID)
)
|| (
REPLACENULL(FirstName, "") != REPLACENULL(FirstName_2, "")
)
|| (
REPLACENULL(LastName, "") != REPLACENULL(LastName_2, "")
)
|| (
(
Address != Address_2
&& (!ISNULL(Address))
&& (!ISNULL(Address_2))
)
|| ISNULL(Address) != ISNULL(Address_2)
)
This expression is true if the fields do not match with account for NULL values for the same EmployeeID value. And it is true if there is no match for the EmployeeID value from one table for the EmployeeID value in the other table, that is, if there are no rows in both tables that have the EmployeeID value.
You can obtain a similar result in the form of selection using the following T-SQL query:
SELECT
e1.[EmployeeID] AS [EmployeeID],
e2.[EmployeeID] AS [EmployeeID_2],
e1.[FirstName] AS [FirstName],
e2.[FirstName] AS [FirstName_2],
e1.[LastName] AS [LastName],
e2.[LastName] AS [LastName_2],
e1.[Address] AS [Address],
e2.[Address] AS [Address_2]
FROM
[JobEmpl].[dbo].[Employee] AS e1
FULL OUTER JOIN [JobEmplDB].[dbo].[Employee] AS e2 ON e1.[EmployeeID] = e2.[EmployeeID]
WHERE
(e1.[EmployeeID] IS NULL)
OR (e2.[EmployeeID] IS NULL)
OR (COALESCE(e1.[FirstName], N'') <> COALESCE(e2.[FirstName], N''))
OR (COALESCE(e1.[LastName], N'') <> COALESCE(e2.[LastName], N''))
OR (COALESCE(e1.[Address], N'') <> COALESCE(e2.[Address], N''));
Now, let us connect the elements "Merge Join" and "Conditional Split" as follows:
Img.36. Connecting "Merge Join" and "Conditional Split" elements
Next, we create an OLE DB destination element in the following way:
Img.37. Creating a destination element
Now, we map the columns:
Img.38. Columns mapping
We set "Error Output" tab by default:
Img.39. "Error Output" Tab-2
We can now join "Conditional Split" and "OLE DB JobEmplDiff" elements. As a result, we get a complete data flow:
Img.40. A complete data flow
Let us run the package that we have obtained:
Img.41. Work of data flow
Upon successful completion of the package work, all its elements turn into green circles:
Img.42. A processed package without errors
If an error occurs, it is displayed in the form of a red circle instead of a green one. To resolve any issues, you need to read the log files.
To analyze the data difference, we need to derive the necessary data from the EmployeeDiff table of the JobEmplDiff database:
SELECT
[ID],
[EmployeeID],
[EmployeeID_2],
[FirstName],
[FirstName_2],
[LastName],
[LastName_2],
[Address],
[Address_2]
FROM
[JobEmplDiff].[dbo].[EmployeeDiff]
Img.43. Data difference-1
Here, you can see the Employee table from JobEmpl database, where Address isn’t set, and FirstName and LastName are mixed up in some columns. However, there is a bunch of missing rows in JobEmplDB, which exist in JobEmpl,
Img.44. Data difference-2
Comparing database data with the help of dbForge Data Compare
Let’s make data comparison with the help of dbForge Data Compare for SQL Server, which is also included in SQL Tools.
For this, in SSMS, right-click the JobEmpl database and in the context menu, choose Data Compare\Set as Source
Img.45. Choosing a data source for data comparison
Choose the second database JobEmplDB as Target and click the green arrow between the source and target:
Img.46. Starting the database data comparison setup
In the opened project of database comparison, click the "Next" button:
Img.47. Defining source and target for data comparison
We leave the following settings by default and press the "Next" button:
Img.48. Data comparison options
On the "Mapping" tag, we choose the desired Employee table and click the ellipsis:
Img.49. Choosing the Employee table
The field mapping window opens,
Img.50. Field mapping
In our case, only four fields are mapped, because two last fields are contained only in the JobEmpl database and are absent in the JobEmplDB database.
This setting is quite convenient when the names of columns in the source table and target table do not match.
"Column details" displays the details of the definitions of the columns from 2 tables: on the left – from the source database, on the right – from the target database.
Let’s hit the "OK" button. Note that you can set another key in "Comparison Key" including a customizable one. Now, click the "Compare" button to start comparing data in the databases,
Img.51. Starting data comparison
After that, the progress window will appear:
Img.52. Data comparison progress
Here we can see that 2 047 000 records are missing in the Employee table of the JobEmplDB database, which is present in the Employee table of the JobEmpl database, and the remaining 1000 records differ because they have no Address value in the source and because the FirstName and LastName field values are swapped in the table source. It is also possible to set visibility and sorting of fields:
Img.53. Data difference-3
Img.54. Setting visibility and sorting of columns
There is also a "Find and Replace" button:
Img.55. Selecting the "Find and Replace" button
Img.56. Find and Replace
You can use dbForge Search tool to search by selected data.
Conclusion
To sum up this article, we have studied the possible ways of comparing schemas and data of two databases. To illustrate this, we compared schemas and data of the Employee table in the JobEmpl and JobEmplDB databases. dbForge Schema Compare and Data Compare are Devart tools, included in SQL Tools, that enable us to perform database schema comparison without creating 2 projects for compared databases beforehand. They also enable users to compare database data without creating any extra tables.