File Share Graph Connector for Microsoft Search: Improvements and upcoming featu...

Introduction

With an ever-increasing requirement for data driven decision-making, organizations want to store the information in its entirety. This information is maintained in multiple systems inside an organization and is accessed by multiple users. Because the systems are siloed, the users spend a lot of time searching and gathering the information. Microsoft Graph Connectors empower search administrators to create connectors to various data sources (on-prem and cloud) so that the users can search the data from multiple sources in a single place using Microsoft Search, the workplace search solution.

Microsoft Graph Connectors – An overview

File Share Connector

File Share connector is an out of the box connector from Microsoft. The connector is set up on premise and crawls on-prem files and folders. Because there are firewalls implemented in an on-prem setup, Microsoft has built an on-prem agent – Graph Connector Agent, which runs locally on a machine, crawls the local data and indexes it in the cloud for discoverability. Let us look at some of the salient features of the connector.

Multiple instances

This latest improvement allows you to club all the file paths that require the same configuration and credentials into a single connection as given in the image below. You can configure up to 20 such file paths or instances in a single connection.

Up to 20 file paths can be added in a single connection

Additionally, you can monitor the crawl statistics of all these instances individually.

Connection pane showing crawl statistics from current and previous crawl for 3 different file paths

File Indexing limits

To include/exclude certain files, you can select from an array of properties to apply on the file paths. Some of them are:

Limit by type: Add/remove from various files formats available to be included/excluded for/from crawling.

Limit by days or date: You can choose to index files that are modified within a certain time interval.

Limit by location: If you do not want to index certain files or a location, you can exclude the same using regular expressions

Multiple options for limits for file indexing

Group search content in dedicated verticals

Since the data is spread in multiple locations admins need to create multiple connections to connect to these data sources. With custom verticals, you can club the results from these locations into a single place so that the users can find the relevant content on a single page.

Select from existing connections to include them in the vertical

You can also create a custom vertical for a connection. For example, for all HR policies, you can create a custom vertical, HR Policy, and let the users land on this page for any file search regarding HR policies.

Search results displayed in a custom vertical - “HR Policy”

Content Freshness

After the first full crawl, you can set the frequency of crawls to maintain the freshness of the search index. The crawler identifies the modified files and indexes them in the cloud. Thus, the users can search with the latest index. However, this depends on the number of modified items.

Multiple options to choose from to trigger the incremental crawl

The Road Ahead

We are going to add more features to the File Share connector in the upcoming releases. Soon, for file share connectors you can add more than 5 million items in a single connection provided the quota limit or the tenant limit, whichever is lower, is not crossed. We are also increasing the number of connections from 10 to 30 to give you the flexibility to add multiple connectors. On the search user experience, we are going to enable OCR for your files to let the users search text from images, PDFs etc.

Some of these advanced features will be available for preview soon. Fill this form to express your interest in taking part in the feature preview.

Read more about File Share Graph Connector here:

File share Microsoft Graph connector | Microsoft Docs

Microsoft Graph connector agent | Microsoft Docs

Write to us with questions or feedback - [email protected]