Advanced Search
Experience nearly load-free access to your data
Last updated
Was this helpful?
Experience nearly load-free access to your data
Last updated
Was this helpful?
From May 9th, RealmJoin introduces a revamped search mechanism that enables real-time access to tenant data. This enhancement ensures efficient performance even for tenants with a large volume of user, device, and group objects.
Tables can be searched using the basic or the advanced search feature.
Basic Search uses the Graph API to directly query tenant data, providing real-time results.
Basic Search is always available as a fallback when Advanced Search is not. It also appears during the initial loading of the cached table, allowing users to search immediately—even before the cache is fully built.
RealmJoin's advanced search feature uses caching and tokenization, making searches fast and reliable, and supports Unicode characters like Ø. This tokenization enables partial name searches, so a query like "Ma Mu" can locate a user named "Maren Müller".
Additionally, information not displayed in the table is also searchable, like post-codes or UPN.
Advanced Search allows querying across available table values using a flexible and powerful syntax. Below are the key features and rules:
Tokenized Search:
Search terms are automatically tokenized to improve matching accuracy. For example, typing lu sk will match names like Luke Skywalker. The search always uses a starts-with approach rather than a full-text or "contains" search. This means a search for walk would not return Skywalker.
Logical AND:
All filters are combined using logical AND. Every condition must be met.
Example: Luke Skywalker country:france searches for Luke Skywalker where the country starts with "france".
Column-specific Search:
Use column: to search within a specific column.
Example: country:france searches for entries where the country starts with "france".
Operators:
: → startsWith search
Example: name:jo matches "John", "Joanna", etc.
= → equals search
Example: city=Stuttgart matches exactly "Stuttgart"
!= or !: → negation
Example: city!=Stuttgartor city!:Stuttgart excludes "Suttgart"
Special Behavior:
Umlaut substitution: Umlauts are normalized (e.g., Björn can be found by searching for Bjorn).
Unicode support: Unicode characters are supported in search terms.
Brackets () are ignored in startsWith searches.
Empty value search:
zip:"" searches for empty values
zip: searches for any value
Supported Columns:
Aliases for column names may exist and are listed in brackets
User table:
name, upn, language (lang), jobtitle (job), city, country, zip (plz), enabled, department (dept)
enabled: → accepts true or false (enabled:true)
Groups table
name
Device table
name, operatingsystem (os), manufacturer (manu), model
Get all users of a specific department:
department:"Sales Department"
Get all users with job title CEO:
job:CEO
Get all users with job title CEO not in Germany:
job:CEO country!:Germany
Get all users starting with ADM- with an empty department:
name:ADM- department=""
Get all users with postal code 63065:
zip:63065
List all users with specific domain:
domain.com
Get all Dell XPS devices:
manu:dell model:xps
The table is cached on a schedule, currently set to 1 hour.
In the top right corner of each table, there's a button to refresh the cached data.
The table is available for all users in the same RealmJoin portal instance. Thus, a triggered rebuild will do so for all those users.
The RealmJoin portal has three different instances for load balancing and redundancy reasons.
The rebuild time depends on the size of the tenant and the number of objects. It is not uncommon, that the rebuild might take several minutes. During the rebuild, only the basic search is available. All data is then queried directly via Graph, thus reducing the capabilities in filtering and speed.
