New and Noteworthy

The on-disk format for disk-based Query Tables was updated for release 7.2.0 to improve performance and take into account the new capture data type. Before migrating your application to 7.2, move all files in your Query Table data directory to a backup location. Then proceed as follows, depending on how you use disk-based Query Tables:

The schema for newly created Query Table delta streams was updated. The data field was replaced with the new or old fields. The operation field now reports update events as a single event, instead of the previous insert plus delete events. Existing applications that monitor a Query Table delta stream will continue to work without modification.

Schemas can now extend a parent schema, which must be a named schema in the same module or in a module or interface explicitly imported into the current module. The child schema inherits all fields from the parent, then can extend the parent by adding fields appropriate for the local context. In the child schema, you cannot subtract or modify parent fields, but you can override an abstract parent field declared with a capture field (described below).

In any Edit Schema tab in the Properties view, there is now a Parent Schemas link to show the parent schemas in use. If the current schema is empty or is defined with a private schema, click the down-arrow next to the green plus icon and select Add Parent Schema. In the resulting dialog, select among the named schemas defined in or imported into the current module. In StreamSQL, specify parent schemas with the implements keyword in a create schema statement:

create schema child (foo int, bar int) implements parent1, parent2;

StreamBase Studio was updated to allow parent schema assignments when editing named schemas, table schemas, input streams, Union operators, output streams, Query Tables, and Lock Sets. See Using Parent Schemas for details.

Starting with this release, modules and interfaces can be configured as hygienic, and newly created modules and interfaces now default to this behavior. A hygienic module requires that the schema of input streams for a Module Reference must strictly match the schemas configured inside the module. This has the benefit of eliminating field name conflicts and other difficult-to-debug typecheck errors, and allows for faster typechecking. Hygienic modules that require flexible input schemas must use the new capture fields functionality described next. Modules and interfaces created in previous releases retain their flexible settings by default when opened in release 7.2.0 or later. See Hygienic Modules.

Starting with this release, modules designated as hygienic can be configured to have a capture field on their input stream schemas and table schemas. If a stream connected to a Module Reference contains additional fields not specified in the module's input schema, these additional fields are captured into the capture field, and are then passed around inside the module as an opaque blob. On output from the module, the captured fields are flattened back into the output stream with their original field names and data types.

The capture type is parameterized by a name, and all capture fields in a module with the same capture type name have the same payload type. The scope of the name binding is module-instance, so different instances of a module can capture different sets of fields to the same capture field name.

Input streams and Query Table schemas can contain exactly one capture field, which collects all additional fields. Hierarchical tuples can contain one capture field in each tuple and sub-tuple. Thus, output streams and outgoing table schemas can end up containing more than one capture field to be expanded; all such capture fields are expanded when exiting the module, potentially leading to name conflicts. See Capture Fields.

The EGit Eclipse plug-in provides seamless access to files managed by the Git version control system, including projects hosted on the Github repository. EGit documentation is now included as part of the Studio Help system. EGit joins CVS and Subversive as the three team provider version control plug-ins included with Studio.

Studio can now open EventFlow or StreamSQL files from the Eclipse local history. With a module file open and active, run Window → Show View → Other → Team → History to open the History view. Be aware that when you open a file from local history, StreamBase typechecking is not run on that file, because resources referenced by the historical file may not also be open in Studio. Use a historical view of a module for quick reference.

Studio can also open files from version controlled resources managed by an Eclipse team provider plug-in. With such a plug-in installed and configured for use with your source code repository, Studio views are extended to allow selection from and browsing of the configured repository.

In previous releases, the Studio Compare Editor could only compare EventFlow modules as XML text, and could only compare two modules selected in the Package Explorer or one module against its history in the Eclipse local history view. Release 7.2.0 has extensively reworked the Compare Editor so that you can now compare the state of two EventFlow modules, or two versions of the same module, side-by-side in graphical as well as XML text forms. You can also now compare the state of one module against an earlier version or similar version stored in a supported version control system (CVS, Subversion, or Git).

A tree of components in the top panel is much like the Outline view. When you select the Application node at the top, the comparison by default is two side-by-side EventFlow canvases, with differences marked.

When you double-click an individual component in the top panel's tree, the comparison shifts to the Properties view differences between the two versions of that component. You can select individual tabs of the Properties view on either side of the comparison, and the other side tracks the change. The image below compares the Schema tab of the Query Table in the center of each canvas in the image above.

EventFlow Compare now includes a limited merge capability using the Copy selection from left to right and Copy selection from right to left buttons.

Release 7.2.0 includes layout file loaders for selected source control systems, including CVS and Subversion (accessed through the Subversive plug-in). These loaders handle the loading of an EventFlow module's companion .sblayout file when you ask to load the .sbapp module.

There are many ways to start an EventFlow Compare session for two modules, or a module and an earlier version. These include:

See Comparing EventFlow Modules for further details.

The Properties view for Query operators associated with Query Tables has the following changes:

See Using the Query Operator for further details.

Starting with this release, when you delete an arc entering a Union operator, or you delete the component immediately upstream of such an arc, the corresponding input port is automatically removed from the Union operator, and the count of input ports on the Union Settings tab of the Properties view is decremented.

You can now use parameters in the form ${parameter} for the timestamp values in the Metronome and Heartbeat operators.

The Properties view for Query Table data constructs has a new tab, Initial Contents. You can now specify that, at module startup, a Query Table is empty (the default setting), or is to be loaded from a specified CSV or JSON file, or is to be loaded from CSV or JSON values specified in the tab.

Any CSV loader file you specify must conform to a strict subset of the CSV features supported elsewhere in StreamBase. A JSON loader file must consist of a series of JSON object rows as defined by the JSON standard. If you provide a mismatched CSV or JSON file, or mismatched rows typed in the Initial Contents tab, the typechecking error is shown in associated Query operators, not in the Properties view for the Query Table itself. See Initial Contents Tab for details.

The JDBC Table data construct now has a hyperlink labeled configure sources on the Data Source tab of its Properties view. Use the link to open the project's server configuration file. If no configuration file exists for the current project, clicking this link opens the New StreamBase Server Configuration File dialog to create a new, empty configuration file. See Using the JDBC Table Data Construct.

When using a Query operator associated with a JDBC Table data construct, if a reconnection to the data source becomes necessary, a tuple is now emitted on the Error Output Port of the Query operator with the message Attempt reconnect to data source. When the reconnection succeeds, another tuple on this port reports Success reconnecting to data source.

The EventFlow Editor has a new tab named Annotations. This contains a checkbox to make the current module hygienic, as described above, which is enabled by default for new modules. The new tab also contains per-module typecheck warning settings, described in Typecheck Warnings. See EventFlow Editor Annotations Tab.

The EventFlow Editor's Metadata tab now has a Version field that takes a string value. Use this field to track version numbers for individual EventFlow modules. The value is stored in the EventFlow XML, for use in EventFlow Compare.

Background typechecking, first seen as a preview feature in release 7.1.4, is now enabled as the default for release 7.2.0. Studio now runs all typechecking operations as a background task, with details visible on demand in the Progress view. Dialogs no longer display during typechecking. This feature allows you to enable the Automatically typecheck components preference even when editing large modules with long typecheck requirements. See Background Typechecking.

Studio now maintains a cache of compiled build artifacts when you run a module in Studio, and uses cached items when re-launching the same module. This can speed up launch times by 30% to 50% on subsequent launches. The cache is automatically maintained such that any portions of an application are rebuilt if changed, but unchanged portions are read from the cache. By default, the cache's location is temporary and starts over with each Studio session. Use Window → Preferences → StreamBase Studio → Compiler Cache to specify using a persistent cache directory that maintains the cache between Studio sessions. On the same preference page, there is a Clear Cache Now button. See Compiler Cache Panel.

If an EventFlow Editor session is open for a module, and you delete that module from the Package Explorer view, the Editor session now closes automatically, if there are no unsaved changes pending for that module.

In the Window → Preferences → StreamBase Studio → Authoring panel, there is a new setting, Center canvas on selection during navigation. The setting is selected by default, preserving the behavior of previous releases, which is to auto-center the canvas on the selected component when navigating in two ways:

When you clear this checkbox, Studio highlights the selected component on the canvas, but does not center the canvas. If the component is offscreen when selected, Studio moves just enough of the canvas to make the component visible.

In the Properties view, the Output subtab of the Streams tab now remains active while you edit the current component instead of going blank during edits, as before. The output schema shown in this subtab might become out of sync with the edits you are currently making, especially if you have automatic typechecking disabled. In this case, a message warns that this view is stale. As soon as typechecking succeeds, the Output subtab is updated.

In a View Source dialog, you can now type Ctrl+F to invoke the standard Find/Replace dialog.

In addition to typecheck warning decorations on component icons, typecheck warning text now appears in the Properties view for the affected component. These warnings direct you to the tab and field in error in the Properties view.

Typecheck warning messages were introduced in release 7.1.0, along with the ability to disable categories of warnings for the Studio workspace as a whole using Window → Preferences → StreamBase Studio → Typechecking.

Starting with 7.2.0, you can also specify typecheck warning settings for individual EventFlow modules, or for all modules in a Studio project. For per-module warning settings, use the new Annotations tab of the EventFlow Editor. For per-module warning settings, use the StreamBase → Project Builder panel of the project's Properties dialog. See Typecheck Warnings for an overview, and see the EventFlow Editor Annotations Tab page for warning categories you can control.

Typecheck warning messages seen in the Studio Problems view are now connected to the Eclipse Quick-Fix mechanism. Right-click a warning and select Quick Fix, as described in Typecheck Warning Quick Fixes in the Problems View.

The error stream from the system container now appears by default in the Stream drop-down list of the Application Output view, and is by default a member of the (All Output Streams) list. As before, you can use the select link to include or exclude this stream from the output.

The right-click context menu in both the Application Input and Application Output views now includes Copy as JSON. Invoke this to copy the currently selected input or output tuple to the clipboard as a JSON object. Use this feature to quickly populate the input and expected output tuples for a StreamBase Unit test, using the com.streambase.sb.unittest.JSONTupleMaker class.

Studio has a new Refactor JDBC Operators Schema Wizard that walks through a module and inspects all Query operators connected to JDBC data sources. If the SQL Statement in such a Query operator returns a result set, the wizard changes the operator to use the Explicitly declare fields below option, and populates the SQL Results fields grid based on executing the SQL statement and examining the result set. You can also run the wizard to revert all JDBC Query operators back to the Result set from SQL Query option. See Refactor All JDBC Query Operators for details.

In the Manual Input view, if the schema for a stream has an entry in its Schema Description field, or if there is a description for a named schema used by the stream, hovering over the information icon for the stream shows that description.

When defining the structure of CSV files used as data input for feed simulations, in previous releases, Studio's Data File Options dialog used the single quote as the default quote symbol. Starting with release 7.2.0, the default quote symbol for this dialog is the double quote. See Generation Method: Data File.

In the Package Explorer, a project's subfolders that are in either the module search path or resource search path are now marked with the StreamBase logo as an overlay icon.

The Debug Configuration dialog now includes a Source (Java) tab, which defines the location of source files used to display source when debugging Java code as part of your StreamBase application. By default, these settings are derived from the current project's Java Build Path. You can override the default settings in the Source tab.

Use the General tab of the Properties view for an Extension Point operator to specify a custom icon for that operator to appear on the EventFlow canvas. Icon image files must be in PNG, GIF, or JPG format, and must exist in the project's resource search path. Size your image file as appropriate to fit on the canvas without displacing adjacent operators, aiming for a recommended size of no more than about 64x64 pixels. See Extension Point Operator, and see the ExtensionPoint.sbapp sample in the operator sample group for an example of an extension point with a custom icon.

The sbadmin utility has two new commands:

The three Thomson Reuters RMDS adapters were renamed to reflect the new name of the protocol formerly known as RMDS: Thomson Reuters Enterprise Platform for Real-Time, or TREP-RT. The adapters are otherwise unchanged. See:

The Results grid in the Thomson Reuters Schema Designer view now has an In Summary column, which for each FID row is either blank or contains Yes. Use this column to help select fields when generating a schema for a summary-only tuple. See Thomson Reuters Schema Designer View.

The FIX Schema Designer (FSD), added in releases 7.0.8 and 7.1.3, now supports the FIX/FAST protocol. FAST stands for FIX Adapted for STreaming, and is a FIX variation used by some trading venues. To design a schema to communicate with such a venue, select FAST Data Dictionary File in the FSD's Select Data Dictionary dialog, then navigate to select the dictionary file provided by the venue. In all other ways, the FSD works the same as before.

The input adapters for IBM WebSphere MQ and Wall Street Systems were updated with two new properties: Include Depth Field and Syncpoint on Get. These properties are enabled by default for compatibility with previous releases of these adapters. Disabling both settings significantly improves performance of these adapters. See IBM WebSphere MQ Input Adapter and Wall Street Systems Input Adapter.

The labels for properties of the CitiFX FIX adapter where changed to match current usage in the CitiFX specifications, from Spot to ESP and from RFX to RFS. Existing applications will automatically transition to using the new terms.

StreamBase includes four adapters that take advantage of the third-party WinPcap libraries on Windows, or the libpcap libraries on Linux: the HTTP, IRC, POP3, and SMTP input adapters. Starting with release 7.2.0, these adapters now work with the 64-bit version of StreamBase Server on 64-bit Windows, if you install the latest WinPcap libraries. You no longer need to run the 32-bit Server on 64-bit Windows to use these adapters. On Linux, support is still restricted to StreamBase built for 32-bit Linux platforms.

When developing C++ native code for Windows using any of the supported compilers, specify the new <streambase/Client.hpp> header file in your code instead of the old StreamBaseClient.hpp file. The new file identifies when the Visual C++ 2010 compiler is in use and arranges for the correct, compiler-specific version of the supporting libraries to be automatically loaded. See Configuring Visual C++.

The .NET Client Library was rewritten to improve the performance of creating tuples and accessing tuple field values, with initial tests showing a speedup of 10 to 40 times from the previous API. Most of the .NET API for building StreamBase clients is now a pure .NET implementation, although some classes (such as StreamBaseClient) remain based on the C++ API. (The .NET API for building StreamBase Server monitor applications is unchanged.)

Rewriting most of the API as pure .NET forced some changes in the interface, as described in Migrating .NET Clients in the API Guide. You may need to make source code changes for existing .NET client applications and recompile them in order to accommodate the new API.

Use the provisional new Java interface, TableAccessor, in com.streambase.sb.operator to access Query Tables from user-written Java operators and adapters. This API may change in future releases, but we encourage customers to use it now and report results to StreamBase Technical Support.

In the Java and C++ Client Libraries, use the new readTable() method in the StreamBaseClient class to display all or a specified number of rows from a Query Table or Materialized Window. The .NET API has an analogous new method StreamBaseClient.ReadTable(). The readTable() functionality is also provided in the form of commands for the sbc and jsbc utilities, described above.

The Javadoc for the Schema class now includes a warning that all Schema constructors that take a String name argument should only use null as as the schema name when used by clients, custom operators, and custom adapters. Using a non-null schema name may result in unpredictable behavior in Studio while authoring and typechecking, especially when a named schema exists with the same name. The name argument may be deprecated in a future release.

The hasField(String) method was removed from the Schema class, leaving hasField(CharSequence). This change requires you to recompile your existing Java code.

The getSizeUpperBound() method was removed from the Tuple class. This method was not part of the published API, but if your code uses this method anyway, replace it with getTupleLength() (also an unpublished method) or rewrite the code to remove the need to determine a tuple's size in bytes.

All methods related to ByteOrder, previously marked with @exclude, but visible with Eclipse's auto-completion feature, are now marked as deprecated.

In the .NET Client Library, the StreamBaseClient::SetDynamicVariable() method was renamed as shown here, with an initial uppercase S. The previous spelling with lowercase s was retained for compatibility, but is now deprecated.

In the com.streambase.sb.unittest package, the SBTestEnvironment interface and its implementation BaseTestEnvironment were simplified to avoid a compile time dependency on sbserver.jar. This change break backwards compatibility on these classes and their subclasses. Most subclasses will only need to delete the method createRuntimeEnv().

For release 7.2, StreamBase Studio is based on Eclipse 3.7.1. When installing Studio in the form of plug-ins for a separate copy of Eclipse, you must start with Eclipse 3.6 or 3.7. Eclipse 3.5 is no longer supported for that purpose. See Studio Eclipse Update Site.

For previous StreamBase releases, StreamBase Systems made available a special edition for 32-bit Linux built using the GCC 3.3 compiler, for use with third-party libraries that required using GCC 3.3 or earlier. In particular, this edition supported the privately built Reuters RFA adapter for 32-bit Linux. Advancements in other libraries used by StreamBase make it no longer possible to support the Linux-GCC33 build or the Reuters RFA adapter for 32-bit Linux. Customers using Linux can still subscribe to RMDS 5 feeds using the embedded Thomson Reuters Enterprise Platform for Real-Time Subscribe Adapter (formerly named the Thomson Reuters RMDS Subscribe adapter).

This release adds Microsoft Visual Studio 2010 to the list of supported compilers for building custom C++ clients and .NET clients on Windows. This means StreamBase now supports version 4 of the .NET framework when used with Visual Studio 2010. See Configure Visual Studio 10.0 Projects to Build Clients for details.

In previous releases, the primary StreamBase kit included the DLL and library files required to build C++ clients and custom functions on Windows. Starting with release 7.2.0, those files are now in a separate kit with its own installation package.

To develop C++ native code for Windows, after you install StreamBase, you must now install the Windows Native SDK Package from the installer file provided as a separate download from download.streambase.com. The same kit is on the StreamBase DVD-ROM and on the downloaded ISO image file. In all cases, the installer file's name is in the form StreamBase-windows-native-SDK-release-datecode.msi. There is one installer package for both 32-bit and 64-bit editions of Windows. Development in a .NET language does not require this kit. See Configuring Visual C++.

This release adds support for Vertica 5.0, both for JDBC access and for high-performance bulk loading with native connectivity.

The bundled version of Logback logging was updated from 0.9.26 to 0.9.30. The version of SLF4J is now 1.6.3.

Demonstrates how to use a capture field in the schema of a Query Table to make that table reusable in different copies of its containing module. Each instance of the table holds key-value data with different value data types for different instances. See Capture Fields for Generic Data Store Sample.

Demonstrates how to use capture fields in conjunction with parent schemas to create a reusable module that can match orders for FX trading in one instance and can match orders for Equities trading in another copy of the identical, unchanged module. The module uses abstract schemas for its input stream and Query Table that become concrete schemas in actual use. See Capture Fields and Parent Schemas Sample.

Provides sample Java code that illustrates the new StreamBaseClient.ReadTable() method, which is used to display all or a specified number of rows from a Query Table or Materialized Window.

Provides sample Java code that illustrates the use of methods in the TableAccessor class that lets Java operators and adapters interact with Query Tables. In particular, this class lets you insert and replace rows, read a single row based on the primary index, perform more complex multi-row queries, delete rows, and truncate entire tables.

The client sample was updated to provide a solutions file for Visual Studio 2010, and to use the new header file, as recommended in API Changes above.

The on-disk format for disk-based Query Tables was updated for release 7.1.0 to improve performance, and to plan for more flexible table and row metadata. Before migrating your application to 7.1, move all files in your Query Table data directory to a backup location. Then proceed as follows, depending on how you use disk-based Query Tables:

The schema of the standard StreamBase error tuple was changed to preserve the failing incoming tuple in a more easily parsable form as a string field in JSON object format, rather than a blob field as before. An associated new feature is the parsejson() function that you can use to reconstruct the original failing tuple for further processing. The Error Streams sample was updated to illustrate extracting the failing tuple from the error tuple using parsejson().

This change causes a typecheck error in any module created with an earlier release, when loaded into StreamBase 7.1.0 or later, if the module uses Error Streams or error ports on components.

This release provides a Java system property and environment variable that directs StreamBase Server to emit error tuples using the schema of previous releases. You can set this property to continue working in 7.1.0 or later with a large application in which many modules emit and manipulate error tuples. When you have migrated all modules to take advantage of the new schema, change this property or variable to specify using the new error tuple schema.

You can either set the system property for the JVM that will host StreamBase Server or set the environment variable in the environment in which StreamBase Server will run. The following table shows the settings.

To allow significant improvements to the EventFlow Debugger, the internal format of breakpoints was changed. Before migrating an existing Studio workspace from a previous release to release 7.1:

Starting with release 7.1.0, the JVM hosting StreamBase Server no longer sets the streambase.uri or streambase.sbd.leadership-status system properties. This allows separate server instances hosted on the same JVM (such as can occur when running StreamBase JUnit tests) to respond with separate URIs.

Update any expressions that use systemproperty("streambase.uri") to instead use the new getServerURI() function. To obtain the current leadership status, use the getLeadershipStatus() function.

In Java code, use getRuntimeEnvironment().getURI() to request the server's URI. Do not use the StreambaseURI.fromEnvironment() method in any Java operator or Java function that runs within the server.

StreamBase Studio's default preference setting for newly created projects is to keep EventFlow layout information in separate .sblayout files. Projects migrated from earlier releases, including some sample projects, retain their previously set layout settings.

As before, Studio respects and preserves the layout status it detects for each individual EventFlow application or module. That is:

As before, you can create new layout files for one or more application files using StreamBase → EventFlow Layout → Extract to Layout File. For details, see EventFlow Layout Options.

When using keyboard shortcuts in StreamBase Studio, be aware of the following new behavior. When components are placed onto the canvas with a keyboard shortcut (or by dragging from the Palette view), the component is now selected, with a cell edit window opened for you to immediately name the new component. When using keyboard shortcuts in sequence to add several components at once, you must now press Esc or Enter to close the edit window before entering the next shortcut. For example, in previous releases, the following shortcut sequence adds an input stream, Map operator, and output stream:

With the new default behavior in StreamBase 7.1.x, this sequence now becomes:

As an alternative, you can temporarily disable the Automatically edit names of newly inserted components option in Window → Preferences, StreamBase Studio → Authoring. Then, run your keyboard sequence without Enter keystrokes, and re-enable the preference afterwards.

StreamBase 7.1 adds 64-bit Red Hat Enterprise Linux Server 6 to its list of supported platforms.

The version of Eclipse that serves as a foundation for StreamBase Studio was updated to Eclipse 3.6.2, and retains backward compatibility for Eclipse 3.5 plug-ins. If you add third-party Eclipse plug-ins to Studio, make sure they are compatible with Eclipse 3.5 or 3.6.

The Oracle-Sun JDK installed for private use by StreamBase was updated to release 1.6.0_24. As before, you can override this installed default to use a different supported JDK as described in Using an External JDK. The supported JDK versions are listed in Supported Configurations.

Trial installation kits of StreamBase Developer Edition are now available in both 32-bit and 64-bit versions for both Windows and Linux.

StreamBase now supports warning messages issued during typechecking in addition to the typecheck error messages long supported. Warnings provide guidance on settings that may cause complications, but do not prevent the EventFlow or StreamSQL module from compiling and running. For example, you might see a warning for a Query operator connected to a Query Table that a proposed Read operation will result in a full table scan.

Warning icons appear in the tab for a module that has warnings, on its entry in the Package Explorer view, in hover text over the component that has the warning, and as an overlay for that component's icon on the canvas. Typecheck warnings can also be issued on the console by StreamBase Server at runtime.

You can suppress warnings by category from appearing on the canvas by means of a new Preferences page, Window → Preferences, StreamBase Studio → Typechecking, as described on Studio Preference Settings. The number and type of warnings, and the number of ways to suppress warnings, are both expected to increase in future releases. See Typecheck Warnings for further details.

In StreamBase 7.1+, it is no longer a typecheck error to leave most schemas empty and undefined. This feature is described in a new page in the Authoring Guide, Using Empty Schemas, and empty schemas are referenced on relevant pages throughout the documentation.

Lists of numeric data types now participate in the StreamBase data type coercion rules described in Data Type Coercion. For example, a list(int) and a list(long) will successfully merge as the two inputs of a Union operator, with the list(int) elements automatically promoted to longs. Two list(tuple) will merge if an adequate supertype of both tuple schemas can be found.

Coercion rules are not applied for output streams with declared schemas.

The naming rules for StreamBase identifiers were broadened. StreamBase now supports a syntax for escaping identifier names that contain reserved characters, as described on Identfier Naming Rules. The escaped identifier syntax is #"name", a pound or hash symbol, with the string in double quotes. Escaped identifiers work throughout StreamBase for the names of all components, including schemas, fields, operators, tables, lock sets, modules, containers, and data sources. For example, you can use this syntax to name a field the same as a StreamBase reserved word, such as #"AS". As before, identifier names can include Unicode characters as long as Unicode support is enabled for both Studio and Server, as described in Unicode Support.

You can now specify a where clause for a container connection that specifies a predicate expression that limits the tuples sent over that connection. Since you can make more than one container connection to the same stream, you can now connect to one stream multiple times, with a different filter predicate for each. Container connection where clauses are expressed in any of the following ways:

The sbadmin listConnections command now supports the following options:

StreamBase now avoids having to call at runtime any functions that return the same for every call, and that have no side effects. This improves runtime performance for expressions that do not need to be computed for every tuple.

The CSV output of the sbprofile command now includes statistic lines with prefix G that show garbage collection events for the JVM hosting StreamBase Server. See Profiling for details, and see the StreamBaseMonitor API changes that support this feature.

This release supports a new system property, streambase.low-latency-mode. This provides a single switch that enables several other properties to direct StreamBase Server to run in low latency configuration. Setting this property to true performs the following subordinate actions:

The StreamBase Studio → Authoring page in Studio Preferences has a new option that allows you to avoid full-depth typechecking of Extension Points and Module References that have all output schemas explicitly set. This can greatly speed up typechecking for some large applications.

Studio 7.1 has the following changes to support easier renaming of components and refactoring of names throughout a project, and to create a new interface by extracting its components and definitions from an existing module.

In previous releases, the Refactor → Extract as Module command (in the context menu of the EventFlow canvas) was extracting all parameters and their settings from the parent module to the new child module. The command was updated such that it now extracts only parameters actually used by the components selected for extraction. Parameters with default values in the parent module are copied with the same values to the child module. Parameter settings in the child module are updated to link to the parent module's parameter values, in the form ParamName = ${ParamValue}.

At extraction time, you can now specify the location of the new child module to be extracted. This can be a subfolder of the current project, or another project folder in the module search path. These changes are described on the rewritten Extract as Module page in the Authoring Guide.

For operators that have a user-controlled number of ports (including Split, Union, Filter, Gather, and Pattern), in previous releases, you had to specify a new port in the target operator's Properties view, then go back and drag an arc to the new port (or from the new port, for a Split).

Starting with this release, you can now drag an arc to or from a hover port that appears in gray in the bottom right corner while the mouse hovers over the operator. The hover port for a Filter operator is circled in red in the following illustration:

Drag an arc from the hover port to another component, or drag an arc from another component and drop it onto the hover port, and a new input or output port is created for that operator automatically.

In previous releases, dragging an Extension Point icon to the canvas dropped a new, unconfigured Extension Point for which you still needed to specify the interface that the operator referenced. Similarly, dragging a Module icon to the canvas dropped an unconfigured Module Reference and a reminder dialog.

Starting in 7.1.0, dragging a Module icon to the canvas opens a File Chooser dialog showing all EventFlow and StreamSQL files in the current module search path (essentially duplicating in the dialog what the Module Explorer view shows). Dragging an Extension Point icon to the canvas opens a File Chooser dialog showing all interface files in the module search path. In both cases, select the desired module or interface, and a Module Reference or Extension Point is placed on the canvas, ready for final configuration. See Referencing a Module for the behavior differences between the two methods.

If an input stream's schema has field-level descriptions, the description text for a field can now be seen in the Manual Input view as a tooltip when the cursor is in that field.

This release adds Studio highlight and location support for text searches in StreamBase file types in the Studio workspace initiated by the Eclipse Search → File dialog. After a search, in the Search view, use the up and down arrow buttons (or the Ctrl+. shortcut) to select each found location in sequence. Starting with this release, selecting a found location in the Search view opens the editor for the StreamBase file in which the search phrase was found, and highlights the containing line. For EventFlow files, the component that contains the search phrase is selected and centered on the canvas. This feature is described in a new Authoring Guide page, Text Search in Studio.

When generating trace files for a running application that you will use for tuple tracing in the StreamBase Trace Debugger, you can now specify a regular expression filter string to narrow the tuples collected in the trace file. Specify this regular expression in the Advanced tab of the Launch Configuration dialog for the module of interest.

For long-running typecheck operations, the Progress Information dialog now reports component names as they are being checked, information about a module when typechecking passes into one, and, for JDBC Query operators, the JDBC data source.

The StreamBaseMonitor API was updated to collect garbage collection statistics from a running application, along with the statistics already collected. The CSV output format of the sbprofile command was modified to display garbage collection statistics in output lines with the prefix G. See Profiling.

The StreamBase panel of the Properties dialog for EventFlow and StreamSQL files now shows statistics for the selected file after it has been compiled by StreamBase Server at least once. Statistics are shown for the selected module and all submodules it references. See Application Statistics on the new StreamBase File Properties page.

For subfolders in a Studio project folder in the Package Explorer view, the StreamBase option in the context menu now offers two new sub-menu options: Add to Module Search Path and Add to Resource Search Path. If the selected folder is already in one of those search paths, the Add to is seen as Remove from.

The currently running text message is shown in the information bar at the top of the Editor canvas of a running application now shows the full StreamBase URI, including container name, if not default. See Running Applications in Studio.

In Studio's Properties view, when you first open the Chooser sub-panel of the Functions tab, the list of functions now opens only to the category level. Open individual categories to browse for a function, or used the filter text field to search for a function name or pattern.

When editing Java code in StreamBase Studio in a project to which you have not yet added the StreamBase Client Library to the Java Build Path, invoking the Eclipse Quick Fix options with Ctrl+1 now offers to add the Library for you.

StreamBase Studio now supports file locks imposed by Eclipse team provider plug-ins that implement version control systems.

You can now add a description field to document the purpose of constants defined in the Definitions tab of the EventFlow Editor. The Add Table Schema dialog has a new Description tab in which you can document the purpose of a table schema definition as a whole. (You can still populate the separate Description field in the Schema tab of the same dialog, which documents the purpose only of the schema aspect of a Table Schema definition.) In the Interface Editor, you can now add a description for a stream, independent of the description for the stream's schema.

See Using Table Schemas and EventFlow Editor Definitions Tab for details.

The Advanced tab of the Launch Configuration dialog has a new option. The check box for the Filter trace data using this regular expression when tracing field allows you to specify a regular expression that narrows the tuples stored in trace output being collected in that run. See The Advanced Tab on the Editing Launch Configurations page.

When dragging an arc to an input port of a Module Reference or Extension Point, the tooltip for that port now appears, which allows you to see which input stream in the referenced module that port is mapped to.

The tooltip text for a failing component, shown when the mouse hovers over the component, shows typecheck error and warning text first, before the component's description.

For many releases on Windows, Studio has provided the Open StreamBase Command Prompt menu option in the StreamBase context menu for project folders in the Package Explorer view. Starting with 7.1.0, Studio on Linux now provides an analogous option, Open StreamBase Command Shell.

The Import from StreamBase Component Exchange dialog now provides a check box labeled Download components in the background.

A well-formed URL in any text-based editor in Studio now becomes a clickable hyperlink by holding down the Ctrl key while hovering the cursor over the URL. This applies to Studio text editors for README.txt files, server configuration and deployment files, StreamSQL files, and EventFlow files viewed as text. This also applies to URLs in the Description tab in the EventFlow Editor.

The Outline view now shows different icons to distinguish Query Tables and JDBC Tables. Query operators now have different icons depending on whether they are configured for Read, Write, or Delete operations.

As in previous releases, the keyboard shortcut O Q drops a Query operator on the canvas, but the operator's icon does not show its Read, Write, or Delete overlay decoration until you configure the operator. Release 7.1.0 supports the following new keyboard shortcuts, which drop a Query operator onto the canvas with its overlay decoration already in place. After you connect the newly dropped Query operator to a data construct, the Operation field in the Query Settings tab is preselected.

These shortcuts and many more are described on the newly reorganized and expanded Studio Keyboard Shortcuts page.

The New Java Operator and Adapter Wizards were updated to include Schema in the drop-down list of Property types in the New Operator/Adapter Property dialog invoked by the Add button on the wizard's second page. Select this property type to generate an Edit Schemas tab in the Properties view for your generated operator or adapter. See the table in Property Data Type Determines Generated Control on the Using the Embedded Adapter Wizard page for further details.

In addition, the New Java Operator wizard now automatically refreshes the current project's typecheck environment on completion of the wizard, so that your new operator automatically appears in the Project Operators folder of the Insert Java Operator dialog.

The Query operator when connected to a JDBC Table data source was updated as follows:

In addition, the following performance changes were made when using the JDBC Query operator:

The Gather operator's Output Settings tab in its Properties view was updated to use field output grids and an Additional Expressions grid, like the ones used in the Output Settings tabs of the Map and Query operators. See Using the Gather Operator.

In the Properties view for Aggregate operators, presentation of the settings for existing dimensions was simplified in the Dimensions tab. The same information is shown as before, but now reduced from seven to three columns, with the details of each dimension's settings explained in simple text, instead of in columns that might be empty for a particular predicate. The Edit Dimension dialog did not change, only the presentation of the results in the Dimension tab.

In the Filter operator, the behavior of the Up and Down arrows in the Predicates grid was improved. In previous releases, on the Predicate Settings tab of a Filter operator's Properties view, the up and down arrows allowed you to reorder which predicate applied to which output port.

Starting with 7.1.0, the Up and Down arrows also move the arcs connected to each output port without needing to remove and reconnect them. Thus, for example, if a Filter operator's port 2 is connected to an output stream named BigTradesOnly, you can move the predicate for port 2 to port 1 with the Up arrow, after which port 1 is the port connected to stream BigTradesOnly.

When connecting a data arc from a Query Table to a Module Reference that contains only one Query Table marked as a placeholder table, in the Tables tab of the Module Reference, the placeholder table association is now filled in automatically.

The XML Normalizer operator, previously available as an external operator with a separate installation package, is now included in the base StreamBase package. Drag the Java Operators component to the canvas, and select XML Normalizer Operator from the resulting dialog. As before, the operator does not support list and tuple data types. The operator is described in Using the XML Normalizer Operator and has a sample, described on XML Normalizer Operator Sample.

Starting with release 7.0.0, you can change the default colors and fonts for most elements of the Studio interface in the Studio Preferences dialog, in General → Appearance → Colors and Fonts → StreamBase. This feature was extended in release 7.1 in the following ways:

EventFlow Groups have the following updates in this release:

You can now set the background color for individual Notes independently for each Note in the General tab of the Properties view for the Note. The setting for each note is independent of the default Note background color, which is specified, as before, in the General → Appearance → Colors and Fonts → StreamBase page of Studio Preferences. There is now a keyboard shortcut to place a new Note on the canvas: / (a single slash). On Windows, you can also drag text from another drag-enabled Windows application to the EventFlow canvas, which creates a Note containing that text.

The Microsoft Excel External adapter is now available in both 32-bit and 64-bit editions. The two editions provide identical functionality for their target Excel editions. Each edition has a separate installer, and the two editions cannot coexist on the same PC. If you have the 64-bit edition of Excel 2010, the 64-bit Excel adapter supports connectivity to large-scale StreamBase data.

The 64-bit Excel adapter is only appropriate for use with the 64-bit edition of Office 2010 on 64-bit Windows; having 64-bit Windows and 32-bit Office 2010 is not sufficient. Be aware that Office 2010's default and recommended configuration is to install 32-bit Office 2010, even on 64-bit Windows, to retain maximum compatibility with third-party plug-ins, most of which are still 32-bit. Run File → Help in Excel to determine whether your Excel installation is the 32-bit or 64-bit edition. If you are running 64-bit Excel 2010, then you must install the 64-bit Excel Adapter kit.

The StreamBase to StreamBase input and output adapters were updated as follows: