Initialize Intelligent Global Search

This article describes Intelligent Global Search introduced in Version 6.2.0.

Intelligent Search works with different types of Search Entities; i.e., Metric Insights content

  • Elements
  • Datasets
  • Bursts

The search includes meta-information of Search Entities: Information is searched across names, descriptions, categories, tags but not the data they contain.   The system relies on meta-information when locating Elements and Bursts that meet the search criteria.

For Datasets, additional inforation about columns is defined during the Search Setup process.

The article comprises the following topics:

Overview

Global Search means a search introduced in Version 6.1.0. that uses PHP and MySQL Natural Language Understanding (NLU) search facilities.  The feature is based on Python: Engine that receives a user request and, with the help of NLP lib extracts separate lemmas (base forms of words such as: "sales" -> "sale", "is" -> "be"). As a result, the search is very flexible and produces more results that our previous searches.

To configure Global Search:

  1. Go to [Admin] > Select [System] > [System Variables]
  2. Use [gear] icon to Global Search, assign "Y" value

[Save] and [Commit Changes]

System Variables refers to a list of configuration parameters in the system that allow assigning a particular value (Y or N) to enable or disable a desired feature of the system.

  1. Once enabled, the search field appears on a page's top banner.

To enable Intelligent Search:

  1. Use [gear] icon to Data Analyzer, assign value to "Y"

[Save] and [Commit Changes]

1.1. Search Setup:  Admin > System > Search Setup

Search Setup Editor:

  1. Select one or all types of Search Entities
  2. Add Data Sources to search across External Reports
  3. Specify Dataset to search across metadata
  4. [Save] and [Update All Search Indexes] to commit changes.

1.2. Define Search of a specific Dataset

The system looks for lemmas in Objects (Search Entities themselves and Objects associated with them, such as tags, categories, data sources, segment values, and Dataset columns).

Search Entities group results and order them by their rate determining relevance of each item. This rate depends on a number of "hits" and a weight of a field where they are detected (e.g. a name of a field weights more than a field containing its description).

To add new Dataset to search:

  1. [+Add Dataset]
  2. Select desired Dataset
  3. [Save]
  1. Define the Dataset's:
    • Title Column: specifies the field where the search is performed
    • Link Column: allows attaching links to resources, if any
    • Thumbnail Link Column: identifies the preview to be displayed in the search results
  2. Search Index is created when the search is enabled the first time, but it can be updated by clicking [Update Search Index]
  3. [+Add Dataset Column] adds data to the search
2. Available Search Entities

Below is the list of Search Entities available for Intelligent Search. The list is updated once new Search Entities are activated.

Dataset:

  • Dashboard element
  • Dataset column
  • Dashboard category
  • Plugin connection profile
  • Source database connection
  • External Report reference

Dashboard element:

  • Topic
  • Segment value
  • Dashboard category
  • Dataset
  • Segment
  • External Report reference
  • Plugin connection profile
  • Source database connection
  • Notification schedule distribution
  • Topic

External Report reference:

  • Dataset
  • Dashboard element
  • External filter
  • Plugin connection profile

Notification Schedule distribution:

  • Dashboard element
  • Segment value
  • Slack channel
  • MS Teams channel

Security

 

For Admins, no restrictions apply.

For Power and Regular Users, the search engine analyzes user privileges and permissions, including the access type granted (View or Edit), and limits results to those Search Entities to which the user has at least View Access to:

  • Element with or without Category
  • Category
  • Dimension Value of Element
  • Dimension and Dataset through User Map
  • Multi-Metric
  • Dataset
  • Burst
  • Data Source

To learn more about security and definitions of the system major terms, see Security Overview.

To learn more about group access, see [Group Access] Grant Access (View / Edit) to a Category.

Query Language Filters

The system supports applying filters to the search. A filter in this context refers to one of predefined words followed by a colon. The system interprets the words after the colon as filter values.

List of available filter values:

  • Tag
  • Category
  • Dataset
  • Datasource
  • Owner
  • Certifiedby:
  • Certified
  • Type

For [tag] filter, words placed after the colon to the end of the line (or beginning of a filter name) become filter values:

  • "Daily sales category: devops developers tag: test" finds Search Entities by words "daily sales" within categories containing words "devops" or “developers” and a tag containing the word "test"

For [category] filter, the search result after applying "daily sales category: marketing" shows up exact matches for "daily sales category" with a name "marketing" (such category names as "marketing and finance" are included as well).

[Owner] filter is applied to technical and business owners and data-stewards.

[Certified] filter is a boolean filter. It produces only certified elements (or vice versa) and recognizes such values as:

  • "Yes", "true", "Y", "1" for True
  • "No", "not", "false", "N", "0" for False

The system allows searching by a user name (display_name value) in filters:

  • Daily sales certifiedby: John

For [type] filter with predefined values:

  • Element (elements)
  • Reference (references)
  • Burst (bursts)
  • Dataset (datasets)
  • Dataset_data (datasetdata, dataset data, dataset-data)
  • Content_type
  • Metric (metrics)
  • Multi-metric chart (multi metric, multi-metric chart)
  • Report (internal report)
  • External report
  • External content (other external content)

Values from several words can be written with a white space, underscore or hyphen.

[Type] filter is recognized by predefined values. So, if a value is not recognized, it is used as a set of general words. Their value is parsed to the end of the known value and not to the end of the line:

  • "Daily sales type: metric" > "daily sales" elements with a type “metric”
  • "Daily sales type: external report canada"> "daily sales Canada" with a type "external report"
  • "Daily sales type: unknown" > "daily sales type unknown" (no type filter has been recognized)

If a filter-name is not recognized by the system, the words are used for usual search:

  • "Daily sales tag: support" finds Search Entities by words "daily sales" and shows only those tagged the word "support"
  • "Daily sales t@g: support" finds Search Entities by words "daily sales t@g support"

Query Language Enhancers

Search Engine allows recognizing additional symbols: [plus], [minus] and [double quotes].

  1. [Plus] limits the results to include the specified word or phrase
  2. [Minus] exclude specified items
  3. [Double quotes] applies a special form of a word or  order of word for the search

The enhancers can be also applied to query language filtering:

  • Daily sales category: "finance marketing"
  • Daily sales category: -finance