Chapter 08. Searching

In this chapter, we will explore the following concepts:

• Simple and advanced searching with Webtop

• Saving searches

• Full-text indexing

• Subscription

Locating Objects

The previous chapters showed how to create and modify objects in the repository. We saw that various mechanisms could be used for this purpose including programming and interactive scripts using IAPI or IDQL. However, the most common mechanism of interacting with the repository remains applications, particularly Webtop.

The same can be said about locating documents or, more generally, objects within the repository. Webtop provides one of the easiest available interfaces for accessing content within the repository. Typically, consumers of information are quite business savvy and the alternatives to Webtop for searching documents are less desirable to them.

There are two key ways of locating objects within the repository:

1. 1. Navigating through the browser tree to a known path

2. 2. Searching using the words that may...

Locating Objects

The previous chapters showed how to create and modify objects in the repository. We saw that various mechanisms could be used for this purpose including programming and interactive scripts using IAPI or IDQL. However, the most common mechanism of interacting with the repository remains applications, particularly Webtop.

The same can be said about locating documents or, more generally, objects within the repository. Webtop provides one of the easiest available interfaces for accessing content within the repository. Typically, consumers of information are quite business savvy and the alternatives to Webtop for searching documents are less desirable to them.

There are two key ways of locating objects within the repository:

1. 1. Navigating through the browser tree to a known path

2. 2. Searching using the words that may be found within the metadata or content

The navigation mechanism is used when the user knows (or can guess) where an object is located within the folder hierarchy...

Search Process

Webtop enables searching for objects in two ways and there are some nuances to each approach that we will explore in this chapter. However, there is a common underlying pattern to the search process either way.

Webtop allows the following sequence of steps for searching, though some of these steps are optional (refer to the figure for a better understanding):

1. 1. Specify the search criteria: The search criteria define the conditions that an object has to satisfy to be a part of the results. Search criteria typically include words being searched for in metadata or in content and additional conditions (such as last modification date being later than a specified date) on the metadata. The criteria may be explicitly specified by the user, may be implicit, or could be retrieved from a previously saved search.

2. 2. Submit the search request: The search request is submitted once all the desired criteria have been specified.

   

   Webtop passes on the search request to the Content Server where...

Simple Search

Simple search is simple in terms of what the user has to do to perform the search. There is just one field to specify the search words (criteria) and a button to submit the search request. The biggest benefit of simple search is that the user just specifies the search words and gets to the results quickly. This is how it looks on the screen:

Since version 5.3 of Documentum, objects in multiple repositories can be queried for one search operation. For a simple search, all repositories designated as default are searched. Multiple repositories can be set as default using preferences, as described later in this chapter.

A search operation can query full-text indexes as well as object properties. Full-text indexes capture information about the text contents of documents and enable searching the content as well as the object properties. The full-text indexes are created by the Index Server when it is present as a part of the Documentum installation.

The simple search request is...

Advanced Search

Advanced search provides full flexibility to the user for specifying the search criteria. The user can reach the advanced search screen (shown in the following screenshot) when initiating a new search or when revising the last executed search. The biggest benefit of advanced search is that the user can be very specific about the search criteria and is more likely to get relevant results, particularly when there is a large number of potential matches for the search words.

Advanced search enables the user to be very specific about the search criteria in the following manner:

1. 1. Additional areas can be specified to be included in the search besides default repositories. Including additional areas may require re-authentication against the new areas, if the authentication credentials have not been saved. Login preferences can be used to cache login credentials to avoid re-authentication at such times.

2. 2. Specific locations — folder paths and cabinets can be used for the search...

Interacting with Results

The results from a search request, simple or advanced, are shown in the content pane. The content pane is the area in Webtop other than the header, footer, and left navigation. For practical purposes, the content pane can be considered the main area of the screen.

The result objects are shown as a list in the content pane and pagination is available if the result list size is more than the number of items displayed on one page. When page navigation (shown in the next screenshot) is enabled, users can go from one page to an adjacent page, jump to a specific page, or jump to the first or the last page.

If the size of the result list is very large, Webtop may start showing results before the processing is complete. In this case a message — "Processing…", appears in the header indicating that more results may be arriving.

This message is a hyperlink to a Search Status page. The search status page shows the status of the search request in terms of each source — a repository...

Search Preferences

Preferences in Webtop allow users to store their preferred ways of interacting with the application, so that they can avoid specifying these choices repeatedly. Essentially, preferences help users to work efficiently. Search preferences are the preferences that affect search behavior for a user.

The search preferences allow users to specify the desired search behavior in the following ways:

1. 1. Columns specify the set of properties that are displayed in the search result, as shown in the following screenshot:

   

2. 2. Favorite repositories indicate the repositories a user accesses frequently. New repositories can be added to this list by providing information about new connection brokers. The newly added repositories stay in the Favorites list only for the duration of the current user session.

   In order to keep these additional repositories permanently available even after the session expires, the connection brokers need to be added to the dmcl.ini file on the application server...

Full-Text Indexing

We have already seen that full-text indexing affects the search in significant ways. It is worth understanding the fundamental concepts of full-text indexing even though it is an optional component of the Documentum platform.

Full-text indexing is implemented by the Index Server, which is an optional component. One Index Server can provide indexing for multiple repositories and, thus, multiple Content Servers. An Index Agent is associated with a Content Server and supports the indexing needs of the associated repository.

The Index Server participates both in creation of and searching of the full-text indexes. Full-text index creation is coordinated by the Index Agent through the FAST Index Plugin (see the note about FAST in Documentum Product Notes later in this chapter). Querying the indexes is coordinated by the Content Server using the FAST Query Plugin. This is shown in the following figure:

The Index Server receives requests for indexing when a sysobject undergoes...

Frequently Accessed Objects

Searching and navigation provide convenient ways to locate objects based on criteria or location, respectively. However, if there are certain documents that a user accesses frequently, these approaches are still somewhat inefficient. Webtop provides two mechanisms for accessing such objects quickly — subscriptions and shortcuts.

Subscriptions

Subscriptions represent bookmarks, favorite locations, or favorite documents. Users can subscribe to the objects or paths that they access frequently. All the subscribed objects show up under the Subscriptions node in Webtop.

Objects can be subscribed to or unsubscribed via the Tools | Subscribe and Tools | Unsubscribe menu items in Webtop.

A user can also subscribe to notifications for events on objects. By default the checkin event for an object generates notifications. Other events can also be configured for notification. Notifications can be subscribed and unsubscribed by using Tools | Turn on notification and Tools...

Help—Some DQL Queries

While this chapter focused on searching using Webtop, the DQL SELECT query is also used to perform searches. We have already seen SELECT queries in several chapters, so we will only look at the full-text support in DQL queries here. The following queries assume that an Index Server is present in the Documentum deployment.

While various aspects of full-text searches are supported in DQL, we will look at the following key aspects:

1. 1. The SEARCH clause enables searching the full-text index.

2. 2. The keywords SCORE and SUMMARY can be used in the selected values list:

   a. The SCORE keyword returns the document's relevance ranking as determined by the Index Server. A higher relevance ranking implies a better match. By default, the results are returned in descending order of SCORE.

   b. The SUMMARY keyword returns a summary of each document as determined by the Index Server.

The following query retrieves documents containing the word hiring in their content or metadata:

SELECT object_name...

Documentum Product Notes

The search behavior can be altered by the presence of Enterprise Content Integration Services (ECIS). ECIS is an optional component of the Documentum platform and it allows external sources of information (such as databases, websites, or other enterprise applications) to be searched along with the Documentum repositories.

EMC provides adaptors for various external data and content sources so that they can also be searched through ECIS. For example, one search request can pull results from two repositories, a database, and Google and show all the results together. If ECIS is not installed, the external source options are not available for searching.

Full-text indexing is implemented by the Index Server, which is also an optional component of the Documentum platform. However, the Index Server license is included with the Content Server license. EMC embeds FAST InStream in the Index Server, by default. Typically, the Index Server uses significant CPU and memory resources...

Checkpoint

At this point you should be able to answer the following key questions:

1. 1. What is the difference between simple and advanced searches in Webtop?

2. 2. What is full-text indexing? How is it enabled? What is the impact of the full-text indexing on search behavior?

3. 3. How can sources other than Documentum repositories be searched from within Webtop?

Test Your Understanding

1. 1. A document AprReport.pdf is linked to only one folder — JohnsDocuments. Jane has only NONE permission on JohnsDocuments but BROWSE permission on AprReport.pdf. Which of the following statements are true within Webtop?

   a. Jane can use the browser-tree to navigate to AprReport.pdf

   b. Jane can use simple search to locate AprReport.pdf

   c. Jane can use advanced search to locate AprReport.pdf

   d. None of the above

2. 2. A document AprReport.pdf has the following metadata: subject='money', title='April Report', and keywords[0] ='finance'. The report document itself contains the word'Boston'. The Index Server is not installed. Which of the following statements are true?

   a. A simple search for 'finance' can find AprReport.pdf

   b. A simple search for 'Apr' can find AprReport.pdf

   c. A simple search for 'Money' can find AprReport.pdf

   d. A simple search for 'Boston' can find AprReport.pdf

3. 3. A document AprReport.pdf has the following metadata: subject='money', title='April...