Message Analyzer enables you to create a Union that correlates two or more fields, properties, or annotations, as described in Configuring and Managing Message Analyzer Unions. Creating a Union can help streamline data analysis and can also simplify what might otherwise be cryptic field names that are difficult to work with. You can create a new Union by selecting the New Union item from the Unions drop-down list in the global Message Analyzer Tools menu or by clicking the Unions button on the global Message Analyzer toolbar. Making such a selection causes the Edit Union dialog to display, from where you can provide the input configuration data that is required to create a new Union.
Go To Procedure
To go directly to a procedure that creates a Union, see Create a Union of Two Data Fields. However, you are advised to review the information contained in this section before doing so.
Configuring a Union
To configure a Union of two or more data fields, properties, or annotations, use the following controls in the Edit Union dialog:
Name — in this text box, specify a Union name that is appropriate for your environment. However, note that you cannot prefix a Union name with a number; otherwise, you can follow your own naming conventions. Note that Message Analyzer permits you to specify a Union name that is similar to one of its child (constituent) field names. For example, a Union could be named “command”, even if one of its child field names is “SMB.command”. This is because the result of such a naming, <UnionName>.command, is unique and would not collide with the child field name.
If the Union name that you specify already exists, Message Analyzer displays a message box with the following prompt:
"The name has been used, please use a different name"
Type — consists of a label that indicates the Union data type. It is likely that the fields, properties, or annotations that you combine into a Union will be of different data types. When this is the case, Message Analyzer performs a fundamental conversion to an appropriate type, as described in Supporting Type Conversions for Union Fields, to ensure that an appropriate data type is generated for your Union.
If you add or remove any field, property, or annotation in the Edit Union dialog, the Type label may update to reflect a new Union data type, depending on the field types you are adding to or removing from the Union configuration.
Category — in this combo box, specify the name of the Category in which to place your new Union. You have the option to either create a new Category, which is retained as an editable drop-down list item for future selection, or you can choose an existing Category from the drop-down. Be sure to create or choose a Category that is meaningful for your environment. All new categories that you specify become subcategories that appear under the default My Items top-level category in the Unions drop-down list. However, if you leave the Category combo box blank, the Union will be stored directly under the My Items category.
Add — click this button to display the Field Chooser Tool Window, from where you can select the fields, properties, or annotations that you want to combine as a Union and create a correlation of fields with equivalent meaning. After you locate each field in the Field Chooser window that you want to add to the Union, either double-click the name of the field, property, or annotation; or highlight it and click the Select button in the dialog to add it to the Union configuration. Each field entity that you specify in this manner is added to the Select fields to include list box. To create a Union, you must have at least two fields, properties, or annotations.
If you want to get the most out of this feature when you are combining fields, properties, or annotations into a Union, you should carefully consider the following suggestions:
The added entities essentially have the same functional meaning, otherwise the resulting Union is likely to be irrelevant.
The added entities are optionally of the same data type, to minimize memory consumption (see Supporting Type Conversions for Union Fields).
The added entities come from different data sources rather than the same data source or module, as this is the intended use of a Union.
If you want to create a Union of two fields, one from a trace file and one from a log file, they will need to be file types that Message Analyzer supports and the log file will have to be parsed so that a module node for the log will appear in the Field Chooser window, from where you can select the fields of the log. To parse text logs, Message Analyzer requires an OPN configuration file. Several configuration files are provided with Message Analyzer by default to parse various types of text logs, such as SambaSysLogs, Netlogon logs, IIS logs, Cluster logs, and others.
To initiate the parsing process, you will need to specify a configuration file in the Text Log Configuration drop-down list in the New Session dialog for a Data Retrieval Session (Files tab) and then click the Start button in the dialog. Thereafter, the name of the log will appear as a node in Field Chooser, from where you can select any of the fields that Message Analyzer parsed for a particular log type whenever you are creating a Union.
Without considering data sources, you might specify arbitrary fields that hold unrelated data, in which case Message Analyzer will only use the first field that you configured in the Union — note that this can have unexpected results.
Return value as options — consist of the following:
Single value — this option causes the Union to display a single value, even if the correlation results in multiple values for a particular field. When multiple values exist, Message Analyzer picks a value for the Union to display, based on the first field that you configured in the correlation.
Set of multiple values — this option causes the Union to display a set of values when the correlation results in more than one value for a field. For example, Message Analyzer might display a set of IP Addresses as “set (192.168.1.1, 192.168.1.2)” in the Analysis Grid column that displays the correlated Union values, assuming you add the named Union as a new column in the Analysis Grid viewer with the Field Chooser window.
Remove — click this button to remove an existing field that you highlight in the Select fields to include list box. Note that you can remove fields only one at a time.
Save — click this button to save the Union configuration. At this time, a new Union is created and added to the Unions drop-down list in the Category that you specified.
Note Your new Union will also be added to the root Unions node in the Field Chooser window after a Message Analyzer restart. Note that you will use the Field Chooser window to add your new Union as a column in the Analysis Grid viewer. In addition, the new Union becomes an asset that you can share with others through the Message Analyzer sharing infrastructure, as described in Managing Unions as Shared Items.
Supporting Type Conversions for Union Fields
When you combine two or more fields, properties, or annotations into a Union, it is often the case that such entities are of different data types. For instance, you might be combining
int data types in a Union, which have incompatible range values and require a type conversion for use in the Union. The following describes how Message Analyzer treats compatible and incompatible field types, with respect to type conversions:
No type conversion — compatible field types do not require any type conversion, such as those fields that are of the same data type. This results in the Union taking on the same data type as the correlated fields. For example, a Union with two correlated fields of type
ushort, will display ushort in the Type label of the Edit Union dialog.
Implicit type conversion — Message Analyzer can accommodate for incompatible field types through the use of a transitive implicit type conversion, if an appropriate implicit conversion path exists for one or more of the correlated fields. Basically, a transitive conversion says that if type A can be converted to type B and type B can be converted to type C, then type A can be converted to type C as well. Message Analyzer performs such a transitive type conversion in order to detect the appropriate data type of minimum sufficient value range that can represent all the types in the Union. This minimalistic approach ensures that the Union consumes the least amount of memory possible. In the case of a Union with fields of type
intis the minimum type that has a suitable value range to handle the range of values that can occur in fields of these types in the Union. Therefore, the Union takes
intas its type, which then displays in the Type label in the Edit Union dialog.
The OPN Programming Guide contains a type conversion table in section 2.1.7 that enables you to map the conversion path taken by Message Analyzer in the previously mentioned type conversion. For example, the type
bytecan implicitly convert to type
short, and a
shortcan implicitly convert to an
int, therefore, type
bytecan convert directly to type
Base type conversion — incompatible field types would remain incompatible if there were no implicit conversions to a common type available. Therefore, when this is the case, Message Analyzer converts all field types to a common base type called
anytype is sufficient to handle the value range of any other data type. The Union would then display the
anytype in the Type label of the Edit Union dialog.
Performing Other Operations on Unions
After you successfully create one or more Unions, you can perform the following operation on any Union that exists in the Manage Unions dialog, as described in Modifying Unions:
- Create a Copy — this is the only command available for built-in Unions, given that built-in Union assets cannot be edited.
You can perform the above operation in addition to the following on any Union that exists in the My Items category of the Manage Unions dialog. Note that these commands are not available for the built-in Unions: