Salesforce Docs Portal

Unstructured Data File Formats and Connectors

Data 360 supports multiple file formats and connector types for unstructured data.

Supported File Formats for Unstructured Data

Data 360 supports these file formats for unstructured data and search index configurations.

Text-Based Files
FILE Format File Extension Search indexing notes harmonization notes
PDF .pdf Limited support for content formatted as tables.

These elements are supported:

  • Tables: Plain text only
  • Text: Plain text, icons, images, links, lists, and headers
  • Lists: Simple lists and single indentation lists
  • Orientation: Left to right, top down, and up to two-columned text (partially)

These elements are not supported:

  • Tables: Icons, images, inner tables, lists in cells, multiple tables, merged cells, and deep indentations beyond a single level
  • Text: Code blocks, ligatures (two or more letters joined together as a single symbol), and non-Latin letters
  • Lists: Deep nested lists beyond a single level
  • Orientation: Right to left, rotated documents (non vertical), and text with more than two columns.
Text .txt    
HTML .html Limited support for content formatted as tables. Supported
CSV .csv Each row is a chunk.  
Log .log Each row is a chunk.  
Excel .xlsx Limited support for content formatted as tables.  
PowerPoint .pptx Chunked by heading. Limited support for content formatted as tables.  
Rich text format .rtf Chunked by heading or paragraph. Limited support for content formatted as tables.  
Word .docx Chunked by heading or paragraph. Limited support for content formatted as tables. Supported
Messages .msg Chunked by body and attachments.  
Email .eml Chunked by body, headers, and attachments.  
XML .xml Chunked by XML element tag.  
Markdown .md Chunked by heading or paragraph. Tables are flattened to text. Supported
Note
Note Data 360 content harmonization supports Word, PDF, HTML, and Markdown file formats only from unstructured data model objects (UDMOs), not from standard data model objects (DMOs), including those that have a file attachment UDMO. Other ingested file types are not harmonized and can be viewed in their source platform linked from the Content Viewer.
Audio, Video, and Image Files
FILE Format File Extension SEARCH INDEXING Notes
MPEG .mpg Audio only
MPEG-3 .mp3 Audio only
MPEG-4 .mp4 Audio only
MPGA .mpga Audio only
Ogg .ogg Audio only
Wav .wav Audio only
FLAC .flac Audio only
WebM .webm Audio only
JPEG (Beta) .jpg

Image-based files. Embeddings include extracted text and captions

The maximum size of an image that can be processed is 20 MB.
PNG (Beta) .png

Image-based files. Embeddings include extracted text and captions.

The maximum size of an image that can be processed is 20 MB.
PDF (Beta) .pdf

PDFs with Images.

Maximum size of a PDF that can be processed is 100 MB

Unstructured Data Connectors

Data 360 has several connectors built for ingesting unstructured data from third-party applications or sources. You can also create a data lake object to ingest unstructured data through an existing connector.

Additionally, you can ingest ingest unstructured data from file attachments on Salesforce objects or upload files directly into Agentforce Data Libraries.

Unstructured Data Limits and Guidelines

For limits and guidelines related to supported file types, see Data 360 Limits and Guidelines.

For billing info, see Billing Considerations for Unstructured Data and Search Index.

Monitor Unstructured Data Connectors

For methods of monitoring your unstructured data streams and ingestion progress, see Monitor the Status of Unstructured Connectors Data Ingestion.

See Also