Table of Contents
- What is Metadata?
- Everlaw's Metadata Search Term
- Date and Time Search
- What's the difference between an exact and non-exact metadata search?
- There are some unexpected fields in my dataset. Is there something special about metadata fields in Everlaw?
What is Metadata?
Metadata is accessory data captured about documents. For example, an email file will contain not just the text of the email, but also information about who sent and received the email, when and where the email was sent, and from what domain the email came from. Similarly, a Word file might contain metadata about who authored the text and when the file was last edited, among other information.
There are two main types of metadata fields: text fields and time/date fields.
Everlaw’s Metadata Search Term
You can use metadata search terms to search across metadata values in Everlaw. Metadata search terms are accessible under the Metadata section on the search page.
Every metadata field available on your project will be represented by a search term. Thus, these fields may vary by project. For example, if you have many email files on your project, you will likely see the To, From, and Subject fields among your list of metadata search terms. However, if you do not have any emails on your project, you may not be able to access search terms for these fields. The list of available search terms will also include any alias or user-editable fields that an admin has created on the project.
Certain metadata terms are pinned to your left-hand sidebar. You can click “Show all terms” to view a list of all of the metadata terms available on your project, and you can pin terms so that they are always visible by dragging and dropping them under Metadata.
Many metadata terms are text terms. This means that their values are alphanumeric text, as opposed to simply numbers, or dates and times. If the metadata field is a text field, a drop-down list will appear displaying the top 25 most prevalent values for the field you selected. As you type into the input box, the displayed values will change to match your input. To the right of the values you will see an estimate of the number of documents that match that exact metadata search. You can also select “(No value)” to find documents that do not have any value for your selected metadata field.
While there are certain common options available for metadata fields, they may not be the same for every project. Documents that are uploaded or processed into Everlaw are often accompanied by a load file that provides metadata for the upload, and a load file may not specify every possible metadata field.
If you cannot find a particular metadata field that you believe was provided in the load file, it’s possible that an admin has hidden or renamed the field on the metadata tab of the admin page. If you are unsure about what a metadata field means and cannot find references to it in the Everlaw help material, or if you cannot search a certain metadata option, it’s best to ask an admin on your case.
Date and Time Search
Using the metadata search term, you can search by date, time, and timezones. First, select a date metadata field (e.g. Date, Date Created, Date Sent, etc.). Please note that for emails, the Date Sent field is roughly equivalent to the general Date field for other documents. Next, a selection box will appear, where you can search by Date only, Date and Time, Time only, and No Date/Time.
If you search for Date and Time, it will return searches that meet both those criteria, not either one. For example, let’s say you search for all emails from Nov 3 at 1pm to Nov 7 at 2pm.
The following conditions will be met by returned documents:
- Any document with both Date and Time values between Nov 3 at 1pm and Nov 7 at 2pm will be returned
- No document that only has a date between Nov 3 and Nov 7 or only has a time between 1pm and 2pm
You can also search by a minimum or maximum limit for date and/or time. For example, if you’d like to search for documents on or after a certain date, select the lower limit, and leave the upper limit blank. If you do this with time parameters, the limit left blank will be auto-populated with 12:00 AM UTC.
You can also select timezones in your date search. The default timezone for documents uploaded without a timezone is UTC. If you search by time range, Everlaw will return documents that fall within that time range, regardless of timezone assigned to the document. For example, if your document’s time metadata value is 9:00 PM America/New York, and you search a time range which contains 6:00 PM America/Los Angeles, the document will be returned by your search because the converted time matches the search. If you are searching for Time Only, you will need to specify a timezone “variant”, such as PDT versus PST (daylight savings time dependent).
Finally, you can select the No Date/Time button to search for all documents that do not include either metadata values. If you’d like to search for all documents with any of those values, select No Date/Time, and then negate the search.
What’s the difference between an exact and non-exact metadata search?
An exact search means you want the value for the metadata field youRunning an exact metadata search means that Everlaw will only retrieve documents whose value for that field exactly matches the text string entered in the search term. For example, if you select the From metadata term and conduct an exact search for “John Smith,” your search will only retrieve documents whose From field exactly matches the string “John Smith.”
There are some occasions where non-exact metadata searches may be more appropriate. For example, metadata may sometimes be inconsistently formatted across documents, particularly if you are receiving documents from multiple sources. In other instances, one person may have multiple email addresses with the same root (e.g., firstname.lastname@example.org and email@example.com) that you want to search across. In either of these cases, it can be useful to conduct a non-exact metadata search. You can do this by making sure that the Exact box is left unchecked.
As previously mentioned, an exact search searches for metadata values that exactly match the user-provided string. Let’s take a look at an example to clarify:
If you were to search for “george” with the Exact box (in the far right) unchecked, then the search would return all documents whose “From” field value contains the (complete) word “george". This means that the search would return all the listed results in the screenshot above, and more.
If, instead, you ran the search with the Exact box checked, the search would only return documents whose “From” field value is exactly that metadata value. For example, if you selected “‘Townsend George’ firstname.lastname@example.org>”, the search would only return documents whose “From” field value is exactly “‘Townsend George’ email@example.com>”.
If you select an option from the drop-down list by clicking that option or hitting enter while that option is highlighted, the search will automatically default to an exact search. You can tell that a search is an exact search because the word “Exact” will be appended at the end of the search string.
The examples below show the difference in search results that checking or unchecking “Exact” can produce.
Exact box unchecked.
Exact box unchecked.
As illustrated by our “jsmith*” example above, you can also use wildcard characters when constructing searches for text-based metadata fields. Click here to find out more about wildcard searches.
Finally, just like with the content search term, inputting multiple terms or phrases separated by a space or comma will default to an OR search. This means you'll be searching for documents with metadata field values that match one or more of the inputted terms or phrases.
You can also search across a range of text by adding “TO” (in all uppercase) between a lower bound word and an upper bound word. The example query below will return all documents whose custodian field value is alphabetically between Craig and Nick, inclusive.
If you want to create an exclusive range search (one that excludes the lower and upper bounds, in this case “Craig” and “Nick”), envelop your search in curly brackets, as in the example below:
There are some unexpected fields in my dataset. Is there something special about metadata fields in Everlaw?
When documents are loaded into Everlaw, the system automatically detects semantically-related fields and groups them together under a single name (for example, fields like “Custodian”, “Custdn”, “CUST”, etc. will be grouped into the "Custodian" field). Searching a grouped field will search across all the fields grouped under that particular field name. This makes it much easier to search across documents with synonymous fields since they are now normalized. For more information about grouped and canonical fields, click here.