Manage Case Search Indexes

About Case Indexes

Eclipse includes a powerful text search capability using the dtSearch® search engine (from dtSearch Corp.). Eclipse administrators can configure and maintain the accuracy of this index using the procedures explained in this section.

Indexes themselves are initially created as part of the import process, unless the indexing step was skipped (for example, to save time). In addition, various changes require that the index be rebuilt, such as:

• Adding new fields to a case and importing data into them.

• Changing the indexing option for a field from No to Yes.

• Editing (changing the value of) fields that have the indexing option set.

• Replacing or adding text files.

This page contains the following content:

• Configure a Case Index

• Build/Rebuild an Index

• Monitor Index Processing

• Restart an Index Job

Configure a Case Index

To configure a case index:

1. Get started:

   1. Start Eclipse and log in as an Eclipse administrator.

   2. Open the case for which the index is being configured.

   3. In the Case View pane, click Index Management.

2. Click Default dtSearch Index and index information and options listed on the page and details provided in the next step.

3. Click  and make changes as needed. See the following steps for details on index options.

4. Complete this section as described in the following table.

Index Information

Option	Description
Index Name	The name of the index – do not edit this name.
Index Path	The full path and directory in which the index file is maintained – do not edit this path in Eclipse. To change the index path, use Eclipse Administration. See Basic Case Changes and Actions for details.
Auto Recognize Date, Email, and Credit Cards	When this option is selected, dates, email address, and credit-card number recognition will be used instead of simple character recognition for these items when users search for them. Dates: Looks for anything that appears to be a date (common date formats, including English-language months and month abbreviations). Examples include 15 Feb 2013, 2-15-13, or 2013-02-15. Email: Looks for valid email-address patterns, such as jsmith@myco.com, regardless of alphabet settings for the at sign (@), period (.), and any other punctuation that is part of the email address. Credit card numbers: Looks for numeric sequences that match typical credit card numbers from major credit card issuers, such as 1234-1234-1234-1234. Dashes, spaces, or other punctuation within the credit card number are ignored. NOTE: Index build time may be longer when this option is selected, but search times will typically be faster for these types of searches.
Number of indexes	Originally defined during case creation, the number of index files can be changed here if needed. Particularly for large and diverse data collections, it can speed index and search processing to define multiple index files.

 

5. To automatically rebuild the dtSearch index so that changes are incorporated regularly, select the Auto Indexing Enabled option and one of the following intervals:

Auto Indexing Options

Option	Description
Occurs once at (time)	Rebuild the index automatically (if case changes have occurred) once a day at the specified time. For example, rebuild the index every day at 10 p.m. Enter or click to enter/choose the time.
Occurs every... Starting/Ending at (times)	Rebuild the index automatically (if case changes have occurred) at the specified interval (from 1 to 23 hours) over the specified time period. For example, rebuild the index every two hours between 7 a.m. and 7 p.m. (i.e., when reviewers are at work). Enter the interval (or use the up/down arrows to select) and enter the starting and ending times or click to select them.

 

6. Specify how different characters are classified. To use the default settings, click Load dtSearch Defaults above the Alphabet list. Modify the defaults as described in the following table or skip to step 8 if no modifications are needed.

If changes are made, ensure that each character appears in only one of the sections (for example, if you want an ampersand, &, to be ignored, delete it in the [Spaces] section and add it to the [Ignore] section.)

Note: Alphabet customization applies only to characters in the Unicode range 33-127. The Unicode specification governs other character properties. For details, see www.unicode.org.

Alphabet File

Section	Description
[Letters]	Searchable characters, including all alphabetic characters (upper and lower case) and digits. All letters and digits should be classified as letters (A-Z, a-z, and 0-9). In the Alphabet list, the four entries for each letter include the original letter plus lower case, upper case, and unaccented forms of the letter. Four entries are required for each character.
[Hyphen]	The hyphen character is considered as a searchable character, a space, or is ignored (or all three) depending on the selected Hyphen Option (see step 7). Rarely, other characters might be entered in this section for the same treatment. If you do so, make sure they are removed from other sections (Spaces or Ignored).
[Spaces]	Characters entered in this section are treated as plain spaces. For example, if the period (.) is included in this section, M.D. would be indexed as two separate words, M and D. To make any character in this area searchable, delete it from this area and add it to the [Letters] section. For example, if you want the plus character, +, to be searchable (perhaps you need to find the company name “Green+Smith”), you would remove it from the [Spaces] section and add it as follows in the [Letters] section as shown here.
[Ignored]	Characters entered in this section are ignored. For example, if the period is ignored, the abbreviation M.D. would be indexed as a single term, MD.
[End]	Additional alphabet options; see the next step.

 

7. Additional options are presented at the bottom of the Alphabet list. Except for hyphenation, it is recommended that you do not make changes in this section. CJK options relate to treatment of Chinese, Japanese, and Korean characters.

The following table explains possible hyphen options.

Alphabet [End] Options

Hyphen Option	Description
1	Treat hyphens as spaces. Even-tempered will be indexed as two words, even and tempered.
2	Treat hyphens as searchable characters. Even-tempered will be the indexed term.
3	Ignore hyphens. Eventempered will be the indexed term. This is the default selection.
4	Treat hyphens as all three. This option is rarely beneficial, as many extra words are indexed, often yielding excess and unexpected search results. When this option is selected, each hyphenated pair of words results in six words in the index. The only advantage to option 4 might be where both hyphenated and solid forms of several terms are used and need to be found. For example, a search for email would also include e-mail in the results.

 

8. Noise Words: These are common words that are ignored during searches of your Eclipse database. To use the default noise words, click Load dtSearch Defaults above the Noise Words list. Add or remove words from this list as needed.

9. When finished, click Save. In response to the re-indexing message, click Yes to rebuild the index now or No to do it later as described in the following procedure.

If you select , see Monitor Index Processing.

Build/Rebuild an Index

Complete the following steps to build or rebuild the case index. Rebuilding is needed when the case changes (for example, if additional documents are imported into a case).

1. Get started:

   1. Select an appropriate time to rebuild your indexes.

   2. Start Eclipse and log in as an Eclipse administrator.

   3. Open the case for which relationships are being defined.

   4. In the Case View pane, click Index Management.

2. If users might be logged on, alert them that performance may be affected while you rebuild the indexes. (You may also require that users log off, depending on the size and complexity of your databases.)

3. Click Default dtSearch Index.

4. Click one of the following Index Tools buttons:

   • Full Build: Rebuild the index for all documents in the case database.

   • Incremental Build: Rebuild the index only for documents that have not yet been indexed.

5. Observe progress as described in Monitor Index Processing.

6. If needed, after the index is built, alert users that index maintenance is completed.

Monitor Index Processing

The following paragraphs explain the two types of indexing details that are provided in Eclipse when a case index is built or rebuilt.

Case Management Index Details

After completing Build/Rebuild an Index, view the Index Details and Index Build Status panes in the Case Management tab for index management as shown in the following figure. Click  to ensure details are up to date.

Schedule Manager

View the progress of index processing jobs in the Eclipse Schedule Manager as follows:

1. Get started:

   1. Start Eclipse Administration and log in as an Eclipse administrator.

   2. Click the Eclipse button, , and select Administration.

2. In the System Administration pane, click Schedule Manager.

3. Whenever viewing the Schedule Manager tab, click Refresh periodically in the Job Actions pane to ensure details are up to date; complete of the following steps as needed.

4. Review jobs and statuses: review the list of active jobs in the left pane for an overview of all scheduler activity. Jobs are listed by case name; index jobs include the identifier such as “Index” or “Reindex.”

5. Click  next to a job name to see the individual tasks and statuses. Each index job includes a task (“child job”) for each index file comprising the case index. The jobs list includes color-coding to indicate job and task status:

   Status		Job or Task Is...
   	On Hold	Job/task is waiting to be placed in worker queue.
   	In Queue	Job/task is waiting to begin.
   	In Progress	Job/task is being processed.
   	Complete	Job/task has successfully completed.
   	Failed	Job/task has been stopped due to an error.

6. Find a job/specific details: take any of the following actions to find a specific job or job details:

   1. Click Refresh in the Job Actions pane.

   2. Scroll vertically and/or horizontally in the jobs list.

   3. Click the right border and drag to make it wider.

   4. Sort the list by clicking the sort button, , and choosing from the different options. Click a selected option again to reverse the sort order.

   5. To find specific details in a long list, start typing any of the following items in the search box: the case or agent name, job type, part of the job’s scheduled or completion date, or job status. This is a plain text search that searches these items in order for all jobs (not tasks); the list is filtered as search matches are located.

7. View job/task details: expand a job and click a specific task to view details for the task; click Refresh periodically to ensure details are up to date. Note:

   • When a job is in progress, details and errors for each task display in the Job Details pane when the task is selected.

   • Completed jobs are routinely removed from the list.

   • If a job fails, it remains in the list. Its error log can be saved and the job can be restarted as described in the next procedure.

Restart an Index Job

If an index job fails, take the following steps:

1. Get started:

   1. Start Eclipse Administration and log in as an Eclipse administrator.

   2. Click the Eclipse button, , and select Administration.

2. In the System Administration pane, click Schedule Manager.

3. In the jobs list, expand the failed job and click the task that has failed.

4. Review data in the Job Details pane and/or save details to a separate text file by clicking Save Log and completing the Save As dialog box.

5. Analyze the log and assess other possible issues (such as system resource problems) to determine the cause of the failure.

6. Make needed corrections.

7. After corrections have been made:

   1. Return to the Schedule Manager.

   2. In the jobs list, locate and click the failed task.

   3. Click Reset Job.

8. Monitor the job as described in the previous procedure.