Package com.smartgwt.client.docs

Interface SqlSettings

public interface SqlSettings

SQL Database Settings in server.properties

Although the Admin Console provides a UI to let you to configure database access for DataSources that use Smart GWT's built-in SQL engine, it is also possible to configure these DataSources with manual entries in your server.properties file.

When you manually configure a DataSource like this, you do so by maintaining a set of properties with names structured like this: 

   sql.{dbName}.x.y
where {dbName} is the name of the database configuration you are providing. Note that this database name is just an arbitrary name for a particular database configuration; many of the default ones provided with Smart GWT are named after a database type, in order to make their intended use more immediately obvious, but this is not by any means a requirement.

For the remainder of this discussion, we will assume we are configuring a database with a name of "MyDatabase".

SQL configuration properties

sql.MyDatabase.database.type
This should be set to one of the supported database types. These are:

hsqldbHSQLDB 1.7.x and greater
db2IBM DB2 8.x and greater
db2iSeriesIBM DB2 for iSeries/i5, V5R4 and greater
firebirdsqlFirebird 2.5 and greater
informixInformix 11.5 and greater
sqlserverMicrosoft SQL Server 2000 and greater
mysqlMySQL 3.2.x and greater
mariadbMariaDB 5.1 and greater
oracleOracle 8.0.5, 8i and greater
postgresqlPostgreSQL 7.x and greater
genericA generic SQL92 database, with limitations described in this article

sql.MyDatabase.driver
The name of the JDBC driver implementation. This depends upon your database product and version, and the specific JDBC driver you are using (JDBC drivers can usually be downloaded from your database vendor's website). Bearing in mind the caveat that this information can vary by release and JDBC implementation, here are some suggested values for our supported databases:

hsqldborg.hsqldb.jdbcDriver
db2com.ibm.db2.jcc.DB2DataSource
db2iSeriescom.ibm.as400.access.AS400JDBCDriver
firebirdsqlorg.firebirdsql.jdbc.FBDriver
informixcom.informix.jdbc.IfxDriver
sqlservercom.microsoft.jdbc.sqlserver.SQLServerDriver or com.microsoft.sqlserver.jdbc.SQLServerDriver (Microsoft changed the order of "jdbc" and "sqlserver" between the 2000 and 2005 editions of the product)
mysqlcom.mysql.jdbc.jdbc2.optional.MysqlDataSource
mariadborg.mariadb.jdbc.MariaDbDataSource
oracleoracle.jdbc.driver.OracleDriver
postgresqlorg.postgresql.Driver

sql.MyDatabase.driver.serverName
The name or IP address of the database server

sql.MyDatabase.driver.portNumber
The port on which the database server is listening

sql.MyDatabase.driver.user
The user to connect as

sql.MyDatabase.driver.password
The user's password

sql.MyDatabase.driver.databaseName
The database to connect to. A "database" in this context is a named collection of tables and other database resources that are somehow grouped together by the database product. The specifics of how this is implemented vary by database. Note that some database products use the terms "catalog" or "schema" to refer to the same concept, and Oracle - although it does also have a concept of catalog - uses the term "SID" for this concept.

sql.MyDatabase.interface.type
Indicates how the JDBC connection will be created or looked up; the value of this setting depends on the capabilities of the particular JDBC driver you are using, and is inherently connected to the value of sql.MyDatabase.driver. The following settings are supported:

dataSource - the driver is an instance of javax.sql.DataSource and should be instantiated by Smart GWT Server
driverManager - the driver is an instance of java.sql.DriverManager
jndi - the driver is an instance of javax.sql.DataSource and should be looked up using JNDI
spring - the driver is an instance of javax.sql.DataSource and should be obtained from the Spring context using bean id defined in sql.MyDatabase.spring.dataSourceBean setting. For example:
sql.MyDatabase.database.type: mysql
sql.MyDatabase.interface.type: spring
sql.MyDatabase.spring.dataSourceBean: springDataSourceBeanId

sql.MyDatabase.driver.url
For configurations where sql.MyDatabase.interface.type is "driverManager", this property allows you to manually enter the URL we use to connect to the database. If this property is not provided, we build the URL from other settings such as sql.MyDatabase.driver.serverName and sql.MyDatabase.driver.databaseName.

Other properties
Different JDBC drivers support different properties to support product-specific quirks and features. You can often specify these properties by embedding them as parameters in the URL used to connect to the database.

Alternatively, any subproperty you set on the "driver" in server.properties is applied to the JDBC driver object via Reflection. For example, the MySQL JDBC driver supports a property "useUnicode", which forces the database to use Unicode character encoding. If sql.MyDatabase.driver is com.mysql.jdbc.jdbc2.optional.MysqlDataSource, setting sql.MyDatabase.driver.useUnicode to true means we'll attempt to call setUseUnicode(true) on this class. This would have exactly the same effect as defining the connection URL manually and specifying the parameter useUnicode=true.

Mysql vs MariaDB: there is broad compatibility between these two databases as described in https://mariadb.com/kb/en/library/mariadb-vs-mysql-compatibility/. Within the bounds of that compatibility matrix, you can use database.type 'mysql' and 'mariadb' interchangeably, and likewise with the drivers that you use. However for future compatibility it is recommended that you use database.type: mariadb for MariaDB. This will ensure that as MariaDB implements new features and backompat breaking changes, your application won't run into any gotchas because the Smart GWT server logic will automatically use the right feature set to accomplish documented behavior.

Smartclient properties

sql.mysql.optimizeCaseSensitiveCriteria
This setting affects all MySQL connectors and it is set to true by default. Depending on textMatchStyle case sensitivity in text criteria is achieved by using LIKE BINARY sql comparison operator, which does not use indexed search. Indexes are used with regular "=" comparison operator, which does not ensure case sensitivity. With big amounts of data indexes are critical, so in order to use them and still have case sensitivity supported this setting must be set to true (default). This way we would generate comparison expression like:
<field> = <value> AND <field> LIKE BINARY <value>
where first part ensures efficient indexed search and second part adds case sensitivity to significantly reduced amounts of data. This would be more efficient without indexes as well, cause LIKE BINARY conversion would be performed on less rows anyway.
Setting this property to false would bring back the old behavior, when only LIKE BINARY comparison would be used, which would return same results, but much slower.

sql.MyDatabase.useHavingClause
By default SQL query is generated using traditional "having" clause approach: 

select <selectClause> from ... where ... group by <groupClause> having <groupWhereClause>
Setting sql.MyDatabase.useHavingClause to false makes SQL query use subselect approach when main query becomes subselect and then it is filtered in outer "where" clause: 
select * from (select <selectClause> from ... where ... group by <groupClause>) work where <groupWhereClause>
This may be overridden by setting OperationBinding.useHavingClause.

sql.likeIsCaseSensitive and sql.MyDatabase.likeIsCaseSensitive
The "LIKE" operator in the SQL standard is defined as being case-insensitive, however, some databases default to case-sensitive matching. This is not ultimately not technically in violation of the standard since the standard specifies that the comparison is based on the configured "collation" (the collation is basically the rules for comparing characters and deciding which is first when sorting).

If the default behavior of the LIKE operator is a case-sensitive comparison, then, to achieve a case-insensitive comparison, as required by case insensitive operators such as iEquals, iContains and so forth, the Smart GWT server must generate SQL that uses LOWER() or similar SQL functions to ensure case-insensitive comparison.

In databases where the default behavior of the "LIKE" operator is already case-insensitive comparison, explicit use of the LOWER() function can be slower - certain databases do not realize that this is a "no-op" and perform the query poorly (fail to use indices, etc).

Note that above is true for the equality check as well, because usually databases treat the "LIKE" and "=" operators in the same way, i.e. either they both are case-sensitive or both are case-insensitive.

Smartclient does its best to default this behavior for supported databases, so unless your database is configured to use other than the default "collation", nothing needs to be changed. In case you need to change the default behavior set likeIsCaseSensitive flag as follows:

  • likeIsCaseSensitive: true enables Smartclient to convert both sides of comparison to lower/upper case to ensure case-insensitivity:
    LOWER(table.column) = LOWER('Value')
    and
    LOWER(table.column) LIKE LOWER('%Value%')
  • likeIsCaseSensitive: false enables Smartclient to rely on database case-insensitivity and compare values directly:
    table.column = 'Value'
    and
    table.column LIKE '%Value%'

sql.forceInsensitive and sql.MyDatabase.forceInsensitive
Deprecated. Does the same as "likeIsCaseSensitive" described above, but will be removed soon. Please see "likeIsCaseSensitive" setting docs for the details what it affects exactly.

sql.aliasLengthLimit and sql.MyDatabase.aliasLengthLimit
These properties override the default table alias length limit when using features like DataSourceField.includeVia and DataSourceField.otherFKs. Default alias length limit is set accordingly to the documentation for supported databases and defaults to 128 characters, except these databases:

firebirdsql 63
mysql 256
mariadb 256
oracle automatically set to 128 since DB version 12.2 and to 30 for older versions
postgresql 63

In order to support portability across databases it is advised to keep alias length limit at the lowest supported value. Use global setting sql.aliasLengthLimit to apply limit across all DB drivers, or use DB specific setting sql.MyDatabase.aliasLengthLimit (overrides the global one).

sql.postgresql.useILike
Starting with version 12.0, Smart GWT Server supports the use of a Postgres-specific comparison keyword, ILIKE. This keyword natively does a case-insensitive LIKE, so the Smart GWT driver does not have to do what it normally does to enable this kind of comparison, which is to convert the filter value to lower case and then generate SQL like: 

WHERE LOWER(someField) LIKE 'united%'
When ILIKE is in use, Postgres is able to make use of indexes, which it does not do when we use the "lowercase both sides" strategy, so this is a potentially significant performance enhancer, depending on your application: 
 WHERE someField ILIKE 'United%'

sql.log.queriesSlowerThan
Allows you to specify SQL query execution time threshold in milliseconds (defaults to 10000), which if exceeded query is identified as "slow" and may be logged under specific logging category. See DataSource.logSlowSQL for more details.

See Also: