Difference between revisions of "Modeler:12/RepositorySql"

From CaseTalk Wiki
Jump to: navigation, search
 
(20 intermediate revisions by the same user not shown)
Line 5: Line 5:
The IG and IGG files are mirrored in a SQLite[http://www.sqlite.org] database. These can be queried using sql or viewed by clicking the relevant tables or views.
The IG and IGG files are mirrored in a SQLite[http://www.sqlite.org] database. These can be queried using sql or viewed by clicking the relevant tables or views.


[[File:SqlRepository.png]]
[[File:DataExplorer.png]]


= Functions =
== Toolbar ==
The SQLite function set is extended with a few more functions, especially used inside CaseTalk:


== md5 ==
The buttons from left to right represent the following functions:
* Query functions for save/load and reopening
* Adding and removing IG(G) databases
* Refresh metadata
* Commit changes back to the CaseTalk model
* Run a SQL query
* Navigate back and forth between previous queries
* Configure ODBC connection
* Configure JDBC connection
 
== Metadata ==
 
The various connections to local databases, ODBC and JDBC, are listed on the left panel. Expanding the tree shows tables, views, columns and can be dragged and dropped onto other panels and windows.
 
#Dragging table and column names into the query panels will complement the SQL with the names dragged.
# Dragging columns onto the Repository with OTFTs will create Object Types with roles. Once the initial object type is created, additional columns will become fact types. This is an easy way to manually 'reverse engineer' an existing database schema into a conceptual model.
# Dragging columns into the mapping tab, and onto the lexicalized tree items, will document the source columns as a source to the roles in the conceptual information model for lineage purposes.
 
=== Offline metadata ===
Thanks to the JDBC functionality which outputs catalog information as a text, CaseTalk is not able to parse such output. Hence the catalog content can also be manually crafted or generated using external tools. These files ([[jCatalog|*.jcatalog]]) can be stored in your roaming profile folder under the AppData folder. CaseTalk will pick those up the next time.
 
== Query ==
Queries can be typed by hand or completed using drag and drop of tables, views and columns. Also the parameters are supported. Especially when queries are executed more than once, and only the search parameters are different, thie may help in quickly executing them. CaseTalk will prompt you for parameter values when the query is started.
 
[[File:DataExplorerQueryParameter.png]]
 
The internal databases use a SQLite engine to query the data. These functions are extended with a few more functions:
 
=== md5 ===
The md5 function generates a hash from the value passed as text.
The md5 function generates a hash from the value passed as text.


Line 17: Line 44:
   from otft;
   from otft;


== macro ==
=== macro ===
The macro function supports keywords also defined in the [http://wiki.casetalk.com/Modeler:9/CustomAttr#Keywords custom attribute keywords]. An example query could be:
The macro function supports keywords also defined in the [[Modeler:12/CustomAttr#Keywords|custom attribute keywords]]. An example query could be:


   select  
   select  
Line 24: Line 51:
   from otft;
   from otft;


== soundex ==
=== soundex ===
The soundex function returns a code for the text which allows text comparisons on the sound of them.
The soundex function returns a code for the text which allows text comparisons on the sound of them.


   select otftname, soundex(otftname) from otft
   select otftname, soundex(otftname) from otft


== annotation ==
=== annotation ===
Extract the named annotation from a sectioned comment field. CaseTalk supports Custom Annotations, yet stores them in a single textblob. This function enables sql to extract the named annotation.
Extract the named annotation from a sectioned comment field. CaseTalk supports Custom Annotations, yet stores them in a single textblob. This function enables sql to extract the named annotation.


Line 38: Line 65:
   from otft;
   from otft;


== Registry ==
=== Registry ===
Fetch the registry entry value by the name as a simple function instead of a subselect in sql.
Fetch the registry entry value by the name as a simple function instead of a subselect in sql.


Line 46: Line 73:
   from otft;
   from otft;


== CustomAttribute ==
=== CustomAttribute ===
Fetch custom attribute value using a simple function instead of a long subquery.
Fetch custom attribute value using a simple function instead of a long subquery.


Line 54: Line 81:
   from otft;
   from otft;


= External Database =
=== isValidType ===
To test a value for its datatype definition, you may use this function. The parameters are:
* Value
* Datatype
* Length
* Scale
 
  select
    PassengerId,
    isValidType(PassengerId, 'integer', 0, 0) as validInteger
  from Titanic_Ref;
 
= External Metadata =
 
== ODBC ==
Dragging sqlite databases from the project panel into the database panel will mount them to be queried. Additionally any ODBC connection in the system is listed here. This allows any ODBC source to be integrated in CaseTalk through various queries. To add a new ODBC connection, use the "Shared" button in the toolbar.
Dragging sqlite databases from the project panel into the database panel will mount them to be queried. Additionally any ODBC connection in the system is listed here. This allows any ODBC source to be integrated in CaseTalk through various queries. To add a new ODBC connection, use the "Shared" button in the toolbar.


[[File:ODBC_Add.png|600px]]
[[File:ODBC_Add.png|600px]]


== JDBC ==
If JAVA is installed in your machine (or available in a subfolder of CaseTalk), JDBC connections can be used to connect to external databases as well.
The connection configuration uses *.jdbc files stored in your roaming profile folder and may look like this:
  # Below are Postgresql jdbc connection configurations.
  db.driver.class=org.postgresql.Driver
  db.conn.url=jdbc:postgresql://192.168.1.106:5432/database
  db.username=admin
  db.password=
 
  # Added for external java loading
  db.driver.jar=Z:/Projects/jdbcExplorer/jdbcMetaData/lib/postgresql-42.5.1.jar
== Offline Catalogs ==
The jdbcExplorer outputs catalog information as text for CaseTalk. These texts can be stored elsewhere and imported into your machine to provide the catalog information. These files can also be handcrafted to provide support in using this external metadata in your model. For more information read [[jCatalog]].
== Reference Data ==
If you have downloaded database data in Excel Worksheets, you may import those XLS files to show up as external reference data. (''File\Import\Reference Data'')
== Server Database ==
In case the Enterprise Edition is used, a "Server" database will be listed as well. This is the central repository as used for the CaseTalk Manager and the Teams menu. It enables users with a "System" Account to query the central repository. Handle with caution!
In case the Enterprise Edition is used, a "Server" database will be listed as well. This is the central repository as used for the CaseTalk Manager and the Teams menu. It enables users with a "System" Account to query the central repository. Handle with caution!


= Server Database =
Once logged in a set of usr* views are available to read data to which permissions are granted. It allows queries on the central repository to retrieve results over models and versions.
Once logged in a set of usr* views are available to read data to which permissions are granted. It allows queries on the central repository to retrieve results over models and versions.


[[file:ServerViews11.png]]
[[file:ServerViews11.png]]
== Mapping ==
[[File:DataExplorerMapping.png|600px]]
For lineage purposes the various source schemas can be dragged and dropped on the roles / lexicalized paths of the conceptual model (to the right).

Latest revision as of 10:32, 29 March 2024

Data Explorer

This window (formerly named SQL Repository) enables you to access various databases and execute SQL queries to view the data. The metadata tree on the left contains a list of databases, schemas, tables and views and their columns.

Repository

The IG and IGG files are mirrored in a SQLite[1] database. These can be queried using sql or viewed by clicking the relevant tables or views.

DataExplorer.png

Toolbar

The buttons from left to right represent the following functions:

  • Query functions for save/load and reopening
  • Adding and removing IG(G) databases
  • Refresh metadata
  • Commit changes back to the CaseTalk model
  • Run a SQL query
  • Navigate back and forth between previous queries
  • Configure ODBC connection
  • Configure JDBC connection

Metadata

The various connections to local databases, ODBC and JDBC, are listed on the left panel. Expanding the tree shows tables, views, columns and can be dragged and dropped onto other panels and windows.

  1. Dragging table and column names into the query panels will complement the SQL with the names dragged.
  2. Dragging columns onto the Repository with OTFTs will create Object Types with roles. Once the initial object type is created, additional columns will become fact types. This is an easy way to manually 'reverse engineer' an existing database schema into a conceptual model.
  3. Dragging columns into the mapping tab, and onto the lexicalized tree items, will document the source columns as a source to the roles in the conceptual information model for lineage purposes.

Offline metadata

Thanks to the JDBC functionality which outputs catalog information as a text, CaseTalk is not able to parse such output. Hence the catalog content can also be manually crafted or generated using external tools. These files (*.jcatalog) can be stored in your roaming profile folder under the AppData folder. CaseTalk will pick those up the next time.

Query

Queries can be typed by hand or completed using drag and drop of tables, views and columns. Also the parameters are supported. Especially when queries are executed more than once, and only the search parameters are different, thie may help in quickly executing them. CaseTalk will prompt you for parameter values when the query is started.

DataExplorerQueryParameter.png

The internal databases use a SQLite engine to query the data. These functions are extended with a few more functions:

md5

The md5 function generates a hash from the value passed as text.

 select 
 md5(otftname) as OTFT_HASH, OTFTName
 from otft;

macro

The macro function supports keywords also defined in the custom attribute keywords. An example query could be:

 select 
 macro('%PROJECTNAME%\%IGNAME%\v%IGVERSION%\' || otftname) as FULL_OTFT 
 from otft;

soundex

The soundex function returns a code for the text which allows text comparisons on the sound of them.

 select otftname, soundex(otftname) from otft

annotation

Extract the named annotation from a sectioned comment field. CaseTalk supports Custom Annotations, yet stores them in a single textblob. This function enables sql to extract the named annotation.

 select 
 otftname, 
 annotation('Comment', comment) as comment,
 annotation('Remarks', comment) as remarks
 from otft;

Registry

Fetch the registry entry value by the name as a simple function instead of a subselect in sql.

 select
   otftname as tablename,
   registry('Profile') as IG_Transformation
 from otft;

CustomAttribute

Fetch custom attribute value using a simple function instead of a long subquery.

 select
   otftname,
   customattributes('otft', otftname, 'source', 'modeler') as author
 from otft;

isValidType

To test a value for its datatype definition, you may use this function. The parameters are:

  • Value
  • Datatype
  • Length
  • Scale
 select
   PassengerId,
   isValidType(PassengerId, 'integer', 0, 0) as validInteger
 from Titanic_Ref;

External Metadata

ODBC

Dragging sqlite databases from the project panel into the database panel will mount them to be queried. Additionally any ODBC connection in the system is listed here. This allows any ODBC source to be integrated in CaseTalk through various queries. To add a new ODBC connection, use the "Shared" button in the toolbar.

ODBC Add.png

JDBC

If JAVA is installed in your machine (or available in a subfolder of CaseTalk), JDBC connections can be used to connect to external databases as well.

The connection configuration uses *.jdbc files stored in your roaming profile folder and may look like this:

 # Below are Postgresql jdbc connection configurations.
 db.driver.class=org.postgresql.Driver
 db.conn.url=jdbc:postgresql://192.168.1.106:5432/database
 db.username=admin
 db.password=
 
 # Added for external java loading
 db.driver.jar=Z:/Projects/jdbcExplorer/jdbcMetaData/lib/postgresql-42.5.1.jar

Offline Catalogs

The jdbcExplorer outputs catalog information as text for CaseTalk. These texts can be stored elsewhere and imported into your machine to provide the catalog information. These files can also be handcrafted to provide support in using this external metadata in your model. For more information read jCatalog.

Reference Data

If you have downloaded database data in Excel Worksheets, you may import those XLS files to show up as external reference data. (File\Import\Reference Data)

Server Database

In case the Enterprise Edition is used, a "Server" database will be listed as well. This is the central repository as used for the CaseTalk Manager and the Teams menu. It enables users with a "System" Account to query the central repository. Handle with caution!

Once logged in a set of usr* views are available to read data to which permissions are granted. It allows queries on the central repository to retrieve results over models and versions.

ServerViews11.png

Mapping

DataExplorerMapping.png

For lineage purposes the various source schemas can be dragged and dropped on the roles / lexicalized paths of the conceptual model (to the right).