Property Types: Usage, Examples, and Best Practices
A comprehensive guide to property types, their definitions, use cases, practical examples, and best practices for effective data extraction and automation.
This document outlines the different property types available for automation and data extraction workflows, with guidance on when to use each, practical examples, and best practices.
Property Type | When to use this | Example | Best practice |
---|---|---|---|
Text | To extract free-form text (short or long form). | Extracting company name, or company background | Use for unstructured data. Apply validation or post-processing for expected formats if needed. |
File | To provide a file (e.g., PDF) as a source for other property types. | CIM document | Store original files for traceability. Reference files in downstream processes for auditability. |
Single Select | To create a structured field with pre-defined values (one value only). | Triage outcome: pass or fail | Limit options to mutually exclusive values. Use for status, category, or outcome fields. |
Multi Select | To create a structured field with pre-defined values (multiple values allowed). | Industry: SaaS and Healthcare | Use for tags, categories, or attributes where multiple selections are valid. |
Number | To extract or store numeric values. | Deal size, number of employees | Validate number ranges and formats. Use for quantitative analysis and filtering. |
URL | To extract or store web links. | Company website URL | Validate URL format. Use for references, sources, or external resources. |
JSON | To store or extract structured data in JSON format. | Extracted financials as a JSON object | Use for complex, nested, or dynamic data structures. Validate JSON schema where possible. |
Collection | To create structured tables of related information. Ideal for tabular data. | Table of management team members and their roles | Define clear schemas. Use for repeatable, structured data within documents. |
Reference | To create a link between multiple agents or entities. | Linking extracted companies to company analysis agent | Use for relational data. Maintain referential integrity between entities. |
Page Splitter | To split documents into logical sections or pages for processing. | Splitting a CIM into executive summary, financials | Use to improve processing granularity. Apply when different sections require different handling. |
Updated 24 days ago