Case Text Searching
Emplifi Agent allows you to find specific words in case text associated with cases or issues. This feature is available in the following reports: Case Listing, Frequency, and Chart. The feature must first be enabled for the database and column. For details, see the Case Text Search Setup section below.
To use the feature, add a filter that references case text (GF4 category id) as the category and "CONTAINS" as the operator. The Value, From column on the filter will indicate the words to search for and logic to use when searching.
Case Text Search Setup
Case Text Search relies on a database specific feature that is enabled for use within Emplifi Agent. This allows searching and filtering for cases based on certain words in the case text entries. You only need to perform the set up once per database (production or training).
MS SQL Server Database
Steps to Enable the Searching
To set up case text searching, complete the following steps:
Launch the Emplifi System Manager on the application server.
Open the database connection to the database to enable the feature (performing this in a training environment first is recommended).
The ID System Manager connects with will need permissions to add columns and indexes, typically this should be the same ID that was able to perform the upgrade processOpen Utilities > System Setup for that database.
Select Create Full Text Index (on case_text.case_text) option (Update System Configuration will be selected by default).
Click the Run Process button on the banner.
Once complete, case text search will be enabled. On SQL Server the process of building the index can take a significant amount of time (see technical notes below). The actual index creation may be spawned in the background while control returns to System Manager. This means until the background process is complete text searching will produce unreliable results.
Technical Notes from Enabling Case Text Search
With over 500,000 rows of case text, there was no noticeable performance impact and simple case text searches produce results in less than two seconds. Note the time to return the results to the client is excluded, because the transfer of two rows and 2,000 rows to the user's computer is greatly impacted by the network and what the user's computer can handle.
SQL Server will create a file directory to store index and word information. That directory uses an average of 15Mb of space for every 100,000 rows and the directory will be created underneath the default SQL Server data location.
The process to build the text searching index was about 25 minutes for 500,000 rows of case text (including background processing). This was on an older machine that was not in use, so it can be affected if this option is enabled during the day and depending on capabilities of the database server. We recommend that this feature be enabled in the training database first to understand the full impact.
MSS SQL Server Syntax
Value, From Column | Results |
|---|---|
cold and fries | Case text must contain both the words “cold” and “fries” anywhere in the same text entry. |
"cold fries" | Case text must contain the phrase “cold fries” in the same text entry (you must include the quotation marks). |
cold or fries | Case text must contain either the word “cold” or the word “fries” anywhere in the same text entry. |
"cold fries" or "warm salad" | Case text must contain either the phrase “cold fries” or “warm salad” in the same text entry. |
"cold*" and "fries" | Case text must contain a word starting with “cold” and the word “fries” anywhere in the same text entry. |
Ignored Words
There are certain words that are ignored when searching even if entered (for example, "the" and "of"). The actual list of words is maintained by the database; refer to your database documentation or Database Administrator for further information.
Issue Joining
There are some considerations when searching Case Text as to how the data is joined with the issue table. These joins affect what information the search can find.
Search, Frequency Report, and Chart Report
There are no join options from the Case Text to the issue when running the search for either report. You are basically performing an Any Issue, Any Text (no join). For more information, see the Case Listing Joins section below. The only difference is that the search will only show one row for the Case ID regardless of the number of Issue, Case Text combinations found.
Case Listing Joins
The case listing allows for 4 types of Issue Text joins.
Any Issue, Any Text (no join): If you are using a Case Text filter, any matches on the case text will display the information listed in the Columns view. If you are displaying issue information and the case has multiple Issue rows then each issue will be listed for the number of Case Texts that matched. If you have three case texts on a case and three issues and two of the case texts match, you will get six rows in the report. If additionally have an Issue filter your results will still show Issues multiple times if the case text has multiple matches. If one of the issues from the case match, along with two case text matches (regardless of the case text issue sequence), you will get two rows.
Issue and Text Both (equi-join): When using a Case Text filter the Case Text Issue Sequence will be used to link to the Issue exactly. If using both a Case Text filter and an Issue filter, the only rows that will be included are
if the Case Text matches the entered filter
if the Issue matches the entered filter
if the matching Case Text's Issue Sequence equals the matching Issue's Sequence.
If the Case Text's Issue Sequence is (No Issue) then the Case Text will never be included.
Issue with Optional text (outer join Text): When using a Case Text filter the join will behave as an Issue and Text Both (equi-join) as the filter is forcing the system to include Case Text.
Text with Optional Issue (outer join Issue): When using a Case Text filter any matches on the Case Text will display the information listed in the Columns view. If you are displaying Issue information and the Case has multiple Issue rows then each Issue will be listed for Case Text's Issue Sequence that matched. If the Case Text's Issue Sequence is (No Issue) then the Issue information will be suppressed but the row included in the results and any other information (Case, Case Text) will be shown. If you have three Case Texts on a Case and three Issues and two of the Case Texts match you will get two rows in the report. Issue information will be displayed for those Case Text matches that have an Issue Sequence. If additionally you have an Issue filter your results will behave as an Issue and Text Both (equi-join) as the filter is forcing the system to include Issues.