Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 29 Next »

Queries allow you to find the necessary ones in the content.

GEODI Query language is as simple as Google's. You may just type the words and see the result. But there are more. Semantic queries help you address data much sharper when the words are insufficient.

GEODI queries by date, not just the document date but also dates in the content. These dates may have been written in many different formats but are all the same for GEODI. Dates are semantic entities for GEODI, and GEODI has many ways to query them.

Money is another example. The Discovery tool catches money phrases written in many different ways. So all phrases become one based on the value and the currency. It would be orherwise very diffcult to find them just with word search. Recognizng Money is very useful for auto classification, fine-tuning permission schemas, and more.

There are more semantic features that help you find any document faster and more accurately. This page and the sub-pages will explain all the rules in detail.

Some rules may be complicated or hard to remember, but there are ways to save them as templates and reuse them.

Keep this page open in your browser. (smile)

GEODI queries can be viewed as a list, map, calendar, panel, or report, whichever is best for you. You can switch between views and narrow the query. The views allow you to see different faces of the data and are good ways of getting insight.

Basic Search

  • All data sources, files, social media, or DBs are searched unless you limit the target.

  • We use the term “CONTENT” to address all types. Content is a file, tweet, or row in a database.

  • You may search for any word(s). Search is case-insensitive; that is, the upper or lower case does not matter. Georgia or GEORGIA is the same.

  • if you want to search words with next to each other, put an “ around to search for an exact phrase.

  • You may use “*” at the beginning or end of a word for an unknown part.

  • Use “-” → minus reverses any rule.

  • Use “~” if you are not sure about the spelling.

  • Date search → Just write the date or a date range. Because of semantic search abilities, GEODI finds all content no matter how the date is written.

  • Note search → To search your own or others' notes, just use note or note owner:me

  • You may use any rules together.

  • There are more rules to use. Some are available in the facet area on the left to add more criteria about content types, sources, and recognizers.

Georgia -> Any content that contains Georgia

Geordia Aquarium -> Content with Georgia and Aquarium. These words must exist but not necessarily next to each other.

"Georgia Aquarium" -> Content with Georgia Aquarium. The words must exist next to each other(because of ")

Georg* -> Content with Georgia or George or ..

*City -> Content with smartcity or capacity or velocity (among the first 20)

-Georgia -> Content w/o Georgia

-*City -> Content w/o smartcity or capacity or velocity

Berlon~ -> Berlon~ finds Berlin, Terlon..

January 1st, 2020 -> finds all content with that date no matter how it is written. Jan 1, 2020, 1 Enero 2020, 01.01.2020 or other languages.

1-5 Jan 2020 -> Finds all content fall into this range. 4 jan 2020 for example.

note -> all notes I am allowed to see, you can take notes for all content types and you can open the main content by clicking the note.

note owner:me -> my notes

owner:me -> content added with drag and drop by me.

Search by Content Names&Types

The search rules are valid for names and the body of any content. If you want to limit the search to names, use doc: or content: prefix.

The name means may vary depending on the content type. The name is what you see in the GEODI search result.

GEODI displays content types on the facet. You may always use the types to limit the search to a certain type.

 By Content Name&Type

content:*.pdf → search PDF files,(actually means names ending with .pdf)

doc:Training* → Any content that name starts with Training

doc:*.tweet → Tweets

doc:*.tweet possible -> Tweets with word possible in name or body

doc:(*.tweet possible) -> Tweets with word possible in name only

doc:*.geodinote → notes

Note → is a shortcut for doc:*.geodinote

Search by Dates

Write dates in a natural way. GEODI finds all dates, no matter how they are written.

GEODI automatically recognizes all dates and date ranges in the content and presents them in a CALENDAR view. If you install the language package, it can do for Russian, Arabic, Georgian, Hebrew and other languages as well. The calendar view will give you an idea about the temporal distribution of a query. Times when some news is concentrated, events like the orange blossom festival, will make it easier to perceive temporal events.

contentdate: prefix limits searches to contentdate(like, file date or tweet date),

 Search by Dates

23 April 1920 -> Content with 23 April 1923, 04.23.1923 or ..

1-5 October 1923 -> Any content that has date within the range

May-December 1923 -> 1st of May to December 31 of 1923

contentdate:(23 April 1920) → content that have the given date.

contentdate:today -> the content added today

9.2023-24.09.2023 and date between.

yesterday , today , last week , last month , nest year , tomorrow , next week , this week , previous week , this month , next month

 Search by Relative Date Clauses

You can use the following clauses in queries. The reference date (today) is the date on your computer.

GEODI also recognizes these words in emails and social media(contentdate is exactly known source)

  1. Yesterday, Today, Tomorrow works as expected.

  2. This Week, Last Week, Next Week : These queries are converted to dates from Monday to Sunday.

    • Current Date -> 28.09.2023

    • This Week-> 25.09.2023-01.10.2023 and date between.

  3. This Month, Last Month, Next Month: These queries are converted to dates from the first to the last day of the month.

    • Current Date ->EylĂŒl(09)

    • Last Month-> 01.08.2023-31.08.2023 and date between.

  4. This Year, Last Year, Next year: These queries are converted to dates from 01 January to 31 December of the year.

    • Current Date -> 2023

    • Next year:-> 01.01.2024-31.12.2024 and date between.

  5. Next Decade: When this query is made, it returns the dates of the current week and the 10 years after this week.

    • Current Date -> 02.10.2023

    • Date found on the document(s)-> 02.10.2023-08.10.2023 week, 01.11.2023 and after.

  6. Last Decade: When this query is made, it returns the dates of the current week and the 10 years before this week.

    • Current Date -> 02.10.2023

    • Date found on the document(s)-> 02.10.2023 - 08.10.2023 and before.

The Word Proximity Rule

"()" at the end of the rule affects maps, keywords, and graph pages. This operator is a great insight tool.

  • On the MAP page Accident() means show places near to word Accident.

  • “Accident() Insurance” means finding documents with Accident and Insurance but they should be close to each other.

  • when you use the operator in the keyword page, only related words remain. Places related to accidents or people related to accidents may be found this way.

 Word Proximity Rule

Accident() -> Default distance is 10

Accident(20) -> You may override the default

Accident() Driver() -> They should all be within the dƟstance

*Accident() -> You may use * or tilda(~) operator with proximity , Accident*() , Accident~()

“Driver License UK" -> Quotation rule applies

December 2021() -> Proximity rule works with other feature

Other Rules

 Versioning and Search

ver:last (last version)

ver:0 (old versions)

ver:<version> (a specific version)

ver:[prev | hasprev] (one older version)

GEODI automatically versions content without requiring any settings. By default, all versions are included in your queries. To limit this, you should use the “ver:” keyword.

GEODI does not store previous versions as files. To do this, you can activate “backup” process on a source basis.

 Find Similars

similar:(doc:a.pdf)

similar:”Georgia Aquarium" (finds similar document containig the words)

GEODI finds similarities between text and image contents. It lists similar documents or texts based on the input you provide.

Copied and similar documents are also shown within the GEODI search interface and viewers.

In similarity search, you can also use the following expressions:

  • maxcount:<n> - limit similar count by n.

  • minsimilarity:0.7 - set similarity. The default is 0.7.

  • excludeDuplicates:true exclude copies default =false means copies are listed under similar.

 Find Copies

duplicate (content with copies)

-duplicate (content w/o any copy. i.e unique ones)

duplicate:(doc:a.pdf) (finds copies of a.pdf)

duplicate:"Georgia Aquarium" (finds copy document containg the words)

Typically, 40% of the documents in an organization are duplicates. Duplicates cause confusion and make searches difficult. GEODI finds these and helps you eliminate them.

Typing "duplicate" will find all documents that have copies. Using "-duplicate" will find those without copies.

Duplicate and similar documents are also shown in the GEODI search interface and viewers.

 Search by Source&Parent

source:<>

parent:Georgia* (files whose parent(ZIP or Folder..) has starts with Georgia)

parent:*.ZIP (files in ZIP files)

source:Archive parent:*.ZIP georgia (get files from Archive source, in ZIPs and containing word Georgia)

source:”Archive,HR” (2 sources)

source:”Archive\HR” (HR in Archive, Tree format)

child:”eligible”

You may limit the search results by content sources you have used in the project settings. You may also set rules for parent documents.

You may specify more than one source. If you set source names in a tree format, you should refer the same way.

parent: Search by parent content. ZIPs or folders may have other content under them. if you take a note, then it becomes a child of the content.

child: search by child content. e.g. search files whose note contains “eligible”.

 Search by Image

Drag and drop an image into the search box and find similar ones. You may use a machine part or a wallpaper sample.

(info) This feature must be activated in the project.

 Find by Recognizer

layer:Date

layer:”Name Recognizerı”

layer:”Date Hashtag”

You may limit the results by layer. Find content containing hashtags, dates, a city, a street, or a name. These semantic queries are also used in Classification.

 Find Geo Only

geom

You would query your contents containing location-related information using the keyword “geom”. This could include place names like Georgia, London, parcel numbers, known area codes of phone numbers, etc. These pieces of information contain location references and are visible on the GEODI map.

 Find by Content Language

lang:EN

lang:ES

You can understand which document is in which language. For this, the language recognizer must be added to your project.

To query the language, you need to add the Language Recognizer to your project.

 Search On DB Colums

fieldname:<criteria>

GEODI also searches in databases or structured excel files. You can limit the search by field name. If more than one table in multiple databases contains the same field, the search includes all. So GEODI effectively merges all structured data along with unstructured ones. You can specify the table name for a specific table using ‘parent:’

GEODI collects field names by itself. Indexing and other phases are all auto, but you may use SQL to define how a DB and Tables will be indexed.

PCs discovered by GDE and some other tools may also have a field name that can be searched in this way.

 Query by Content Creator

owner:<username>

The content summary contains the owner, who added the content by drag-and-drop. The owner: query results in the content added by the specified user.

“me” may be used as a username for the current user. So if this query is used in a panel, every user sees content added by themselves.

 Query by Domain Name

Domain:<domain_name>

You can search your website contents in GEODI by the domain name.

 Query by Mail Domain

From:<mail_extension>

To:<mail_extension>

Sort the Search Results

orderby:name

orderby:date

GEODI tries to bring the most relevant content first. But you may define different orders. Sort keywords behave differently in CONTENT and KEYWORDS pages. The following criteria are valid for keywords only.

orderby:layer

Practical Examples

We suggest you to try the following samples with your content. You will see how easy and effective to find content in GEODI.

If you have a rought estimation about time interval then you may specify it in the search phrase. If you are sure that it was a PDF that no problem.

Jan-March 2017 doc:*.PDF Georgia

If the document came via email, the rule remains the same. We specify the PDF just as we specify the email, as both are content types for GEODI.

content:e-Mail Georgia

We know that it was in a ZIP file. The document name was something “Geological..” and contains the the word wall.

parent:*.ZIP doc:Geological* Wall

You looked at hundreds of resumes during an HR process and made notes indicating your opinions. Now you want to find the resumes that received positive notes.
child:(.geodinote “let’s interview”)
child:(.geodinote “not suitable”)

Search for Georgia but not auarium.

Georgia -Aqua*

When you want to search a word and sinonims at the same time you may simply define a GEODI dictionary. The following queries may be identical using a single line dictionary:

Artifical Intelligence

AI

Expert System

Using just words in long documents may be misleading. Lets assume , a 10 page document first word in the first page, second word is in the last page. To prevent this situation you may use proximity rule that is ().

Education() “High School”

find High School within 10 words of education. Try removing () and see the result.

Education “High School”

100..500GBP

Words can be found close to currency expressions that fit the range.

 

Save and Watch Queries

The current query may be saved using the search box menu. Saving enables to reuse of frequent queries and publishing them with other users.

If you watch a query every time, new content or a change happens GEODI will inform you through e-mail. This feature saves a lot of time. You may watch for certain words, but semantic features may be much more useful. Watching a query like “layer:SSN” whenever a document with social security comes you will be informed.  

 

  • No labels