Description
In some cases, it may not be possible to use the canned search options in Clearwell to find documents for review. When this occurs, there is another type of search that can be used. This is called Freeform Search and it is part of the Advanced Search options.
The sections of this document include:
- Logic Operators
- Wildcard Searches
- Advanced Freeform Search Features
- Common Freeform Searches
- Fields for Freeform Searches
To be able to use the Freeform Search, follow these steps:
- Select a case from the ALL CASES dropdown menu.
- Click on ANALYSIS & REVIEW.
- Click on the ADVANCED SEARCH button.
- Next to the STYLE option, select the FREEFORM option.
What this does is change from using KEYWORDS to using QUERIES. All of the regular Advanced Search options are still available to apply to the search being performed. This gives more granular control of the search terms used to find responsive documents. Note that separate text boxes are provided for message queries and file queries. Separate queries are required for messages and files because the data is stored in separate indexes. The query strings entered in each text box are treated as an AND search, along with any other search criteria you specify on the page.
Freeform queries can include terms, fields, and logic operators. The following text fields are specific to Freeform Search.
Field Name | Description | Examples |
Apply this message query: |
Finds messages that match the specified query. |
|
...and apply this file query: |
Finds message attachments and other files that match the specified query. |
Note: The following rules apply to a Freeform Search:
- All searches are case-insensitive.
- Each of the two query fields (message and file) can contain up to 8,000 "tokens". Tokens are individual query elements, such as terms and fields.
- The maximum text length of a query depends on your browser, but is usually 128K.
- In addition to the Freeform Search page, the product supports basic freeform queries in the Basic search and Advanced search Any of these words fields, including phrase, logic operators, grouping, wildcard, and proximity searches.
- Proximity search can only be run within fields, such as u_subject, and u_body, not between fields, as each field is tokenized separately.
- Advanced freeform queries, such as field selection, fuzzy searches and boosting are not supported in the Basic search, or Advanced search in the Any of these words fields.
- Field selection, fuzzy searches and boosting are only supported through the Freeform search screen.
Logic Operators
This table shows the Logic Operators in the Freeform Search that are available for use.
Logic Operator | Description |
+ |
Includes only documents that contain terms after the + symbol (but only the word immediately following the symbol). |
- |
Excludes documents that contain the term after the - symbol. |
OR |
Includes documents that contain either of the terms connected by the OR. The OR operator is the default conjunction operator. This |
AND |
Includes only documents that contain both terms connected by the AND. |
NOT |
Excludes documents that contain the term after the NOT operator. Note that the NOT operator cannot be used with just one term. Like AND searches, NOT searches will treat messages and attachments as separate documents. In the example above, |
Wildcard Searches
There are two types of terms: single terms and phrases:
- A single term is one word, such as "coffee" or "tea." A phrase is a group of words enclosed in double quotation marks, such as "grande latte."
- Multiple terms can be combined together with logic operators to form a more complex query (see "Logic Operators").
The product supports the use of ? and * for single and multiple character wildcard searches, respectively. The table below shows descriptions and examples of the different types of Wildcard Searches.
Single Character Searches |
|
Descriptions of searches that can be performed |
Actual search string |
The single-character wildcard indicates that a match occurs on any character in the wildcard position. For example, to search for "text" or "test," enter: |
te?t |
Multiple-character wildcard searches look for 0 or more characters. For example, to search for "test", "tests" or "tester", enter: You can also use the wildcard searches at the beginning or middle of a term. |
test* |
You can perform wildcard searches in any of the following Advanced search fields: Any of these words, All of these words, phrase, None of these words, Source name and location, Subject, or Attachment/file - Any of the words, and in Basic search. You can use wildcards in phrase and proximity searches in Basic search or Advanced search Any of these words fields. Wildcards in phrase or proximity searches are not supported in any other fields. |
te*t |
Specifically, wildcard queries can be done in Freeform search. However, wildcard searches are not supported when used in phrase or proximity queries. For example, the following query will find hits with "flaming", "flamingo" or "flamingopink" in the body content: |
+u_body:flaming* |
However, the following query will ignore the wildcard and is essentially a simple search for "flaming lawn ornament" and will not find a document with "flamingo lawn ornament" in the body: In this example, the letter "o" completely changes the meaning of the phrase. Note: Searches containing non-ASCII characters and wildcards could return an error due to too many results. If this error occurs, group the non-ASCII characters and wildcards in angle brackets. This prevents the wildcard from running as a separate search. |
+u_body:"flaming* lawn ornament" |
Searches using Grouping |
|
Descriptions of searches that can be performed |
Actual search string |
The product supports using parentheses to group clauses to form sub-queries. This can be very useful if you want to control the Boolean logic for a query. To search for either "coffee" or "tea" and "milk" in a document, use the query: |
(coffee OR tea) AND milk |
Parentheses can also group multiple clauses to a single field. To search for messages that contains both the word "latte" and the phrase "espresso machine" use the query: |
(+latte +"espresso machine") |
Proximity Searches |
|
Descriptions of searches that can be performed |
Actual search string |
Proximity searches find words that have a specific number of intervening words. When performing proximity searches, the word order in the phrase does not matter. The product supports proximity searches containing two or more terms. You can perform a proximity search two ways: |
Both searches will find documents where there are 10 or fewer intervening words between "budget" and "issues" or where there are 10 or fewer intervening words between "issues" and "budget." |
Nested Proximity Searches |
|
Descriptions of searches that can be performed |
Actual search string |
Nested proximity searches combine two query types, proximity and grouping. |
|
Advanced Freeform Search Features
The following types of freeform searches are supported |
||
Fuzzy Searches |
The product supports fuzzy searches based on the Levenshtein Distance, or Edit Distance algorithm. To perform a fuzzy search, add a tilde (~) at the end of a one-word term. For example, to find terms like "foam" and "roams" in the subject of an email, enter the following fuzzy search: u_subject:roam~ |
|
Fields (Click here to see examples) |
Fields let you search specific parts of an email, such as the subject, body, or recipient list. Fields are unstemmed, which means that a match occurs only on the exact text specified in the query. The following table describes the message query fields. |
|
Boosting Terms |
You can boost certain terms in your search relative to other terms. To boost a term, add a caret after the term, followed by a boost factor (a number). The higher the boost factor, the more relevant the term will be considered when ranking results. For example, to search for both "breakfast" and "donuts," but "breakfast" is much more relevant than "donuts," you can enter: breakfast^4 donuts By default, the boost factor of all terms and phrases is 1. Although the boost factor must be positive, you can use a value less than 1 (such as 0.2) to decrease a term's relevance. |
Common Freeform Searches
The following examples show how Freeform Search can be used to satisfy common eDiscovery requests.
Types of searches that can be performed |
Descriptions and Examples |
Finding traffic between two groups |
While the Dashboard can be used to monitor group-to-group communication, in some cases you may want to carry out more detailed searches using the full power of Advanced search. To include a constraint in a Freeform query that restricts the result set to messages that were sent between two groups, use the following query: (sendersDept:"<Group 1>" AND recipientsDepts:"<Group 2>") OR (sendersDept:"<Group 2>" AND recipientsDepts:"<Group 1>") This logic can be made more complex as necessary; such as to track interactions between more than two groups, or to find documents sent from one of several groups to another group. |
Searching for files of a particular type |
Freeform search allows you to distinguish between searches on file content and the file name, so that you can limit your searches to files of a particular type. For example, to find loose XLS files and messages that have XLS attachments that contain the word "budget", use the following file query: +NEAName:(xls) +NEAContent:(budget) |
Finding the blind copy messages that a user received |
In a standard advanced search, the "Recipient" field does not distinguish between the To, Cc, or BCC lines. Using Freeform Search, however, you can easily distinguish between these three fields. For example, to find all messages that were grouped using bcc to someone named Smith, add the following to your message query: +bccListIndexed:(smith) |
Find emails sent between two specific participants |
+((fromListIndexed:("Participant 1") AND toListIndexed:("Participant 2")) OR (fromListIndexed:("Participant 2") AND toListIndexed:("Participant 1"))) The ("Participant #") variable is the Display Name of the participant. |
Fields for Freeform Searches
Below is a list of the Field Names available to use with Freeform Searches. Note: All field names are case sensitive. You must enter all names exactly as shown in the following table.
Fields Available in Message Queries |
|
Field Name |
Description |
fromListIndexed |
The sender of an email. Normally this is a single participant, but in some cases (such as when a message is sent "on behalf of" someone else) there can be multiple senders in the index. |
toListIndexed |
The recipients of the email as specified on the To: line. |
ccListIndexed |
The recipients of the email as specified on the cc: line. |
bccListIndexed |
The recipients of the email as specified on the bcc: line. |
containedSenderListIndexed |
List of senders identified in forwarded emails contained within an original email. |
containedRecipientsListIndexed |
List of recipients identified in forwarded emails contained within an original email. |
ID:<document_ID |
The document ID number of a specific document. For example: ID:0.7.87.2171 |
importance |
The importance of the email. Valid values are:
For example, to search only messages with normal or high importance, add the following to the query: importance: (1 OR 2) |
scope |
The scope of the email. Valid values are:
For example, to search only internal or inbound messages, add the following to the query: scope: (0 OR 1) |
sendersDept |
The group(s) of the senders |
recipientsDepts |
The group(s) of the recipients |
topicNounPhrase |
The most important phrases in the email, as determined by topic classification. Note: This functionality only applies to topics in pre-7.0 cases that upgrade. |
u_subject |
Unstemmed subject. |
u_body |
Unstemmed message text. |
u_quotedTextN |
Unstemmed quoted text regions. |
nonEmailAttachmentNames |
Attachment names found within an email. |
Fields Available in File Queries |
|
Field Name |
Description |
u_NEAContent |
Unstemmed file content. Use this field to find an exact match on the specified file text. |
NEAName |
The filename. |
u_NEAMetadata |
Location where file metadata (such as camera type for a photo) is indexed (in newer versions). |