You can create you’re own databaseType so lets go through how it works.
On the commandline you specify the databaseType using the option
The option can be specified with either [name].properties or just [name]
the .properties will be added if missing. So if you create one, be sure
to have .properties extension.
- The search order is:
- Classpath in schemaspy supplied location
This actually means that if you supply
- It will look for:
- file: $user.dir/my_conf/mydbtype.properties
- Classpath: my_conf/mydbtype.properties
- Classpath: org/schemaspy/types/my_conf/mydbtype.properties
It can contain wast amount of properties so we will break it down. The Properties-file can contain instructions.
extendswhich does what it means, it allows one to override or add properties to an existing databaseType (by specifying a parent/base)
As an example:
which you can see in mysql-socket.properties
include.[n]is a bit different it allows one to add a single property from another databaseType.
[n]is substituted for a number. The value has the form of
As an example:
This would have been valid in the mariadb.properties
Then we have required properties:
- Description for the databaseType (mostly used in logging)
- We will talk more about this one. It’s the connectionUrl used, but it supports token replacement
- FQDN of the JDBC driver as an example
Let’s dive a bit deeper into the connectionSpec.
As an example from mysql-socket:
extends=mysql connectionSpec=jdbc:mysql://<host>/<db>?socketFactory=<socketFactory>&socket=<socket> socketFactory=ClassName of socket factory which must be in your classpath socket=Path To Socket
We mentioned extends earlier.
ConnectionSpec contains the connectionUrl used with the jdbc driver, some might refer to it as the connectionString.
connectionSpec allow token replacement, a token is
In the above example we have host, db, socketFactory, socket.
This means that when used it expects the following commandline arguments:
-h [host] (for host) -db [dbname] (for db) -socketFactory [socketFactory class] -socket [path to socket]
host and db are already known, but
-socket has become a new commandline argument.
The presence of the keys in the databaseType properties file is only for description, it’s printed when
-dbhelp is used as a commandline argument.
(db and host located in databaseType mysql which is extended)
There is also a synthetic token that can be replaced <hostOptionalPort> which combines host and port if port is supplied.
Default separator is
: but can be changed by specifying another under the key
- path to classpath resources that will be used when trying to create the jdbc Driver in java
same as commandline argument
- number of threads that can be used to analyze the database
- regular expression used in conjunction with
-all(and can be command line param
- Which types should be considered tables, default is TABLE
- Which types should be considered views, default is VIEW
Sql query instead of DatabaseMetaData¶
When metadata in JDBC isn’t cutting the mustard. You can replace it with a sql query. They are prepared and supports named parameters as long as they are available. Data is retrieved by column label. So additional columns are ok, but you might need to alias columns so that they are returned correctly to schemaspy.
- alias for :schema
- table that the query relates to (think selectRowCountSql)
- alias for :table
- Possible overrides:
- Fetch comments for a schema, expected columns:
- Fetch comments for a catalog, expected columns:
- Fetch tables, expected columns:
table_name, table_catalog, table_schema, table_comment, table_rows
- Fetch views, expected columns:
view_name, view_catalog, view_schema, view_comment, view_definition
- Fetch indexes, expected columns:
INDEX_NAME, TYPE, NON_UNIQUE, COLUMN_NAME, ASC_OR_DESC
- Fetch table PKs, expected columns:
TABLE_CAT, TABLE_SCHEM, TABLE_NAME, COLUMN_NAME, KEY_SEQ, PK_NAME
- Fetch row count for a table, expected columns:
- Fetch column type for all columns, expected columns:
table_name, column_name, column_type, short_column_type
- Fetch routines, expected columns:
routine_name, routine_type, dtd_identifier, routine_body, routine_definition,sql_data_access, security_type, is_deterministic, routine_comment
- Fetch parameters for routines, expected columns:
specific_name, parameter_name, dtd_identifier, parameter_mode
- Fetch definition for a view, expected columns:
view_definition, text (text has been deprecated)
- Fetch check constraints for all tables, expected columns:
- Fetch ids for all tables, expected columns:
- Fetch ids for all indexes, expected columns:
table_name, index_name, index_id
- Fetch comments for all tables, expected columns:
- Fetch comments for all columns, expected columns:
table_name, column_name, comments