Skip to main content

​
Data Import and Indexing

Populate your search index using push or pull methods to make content searchable.

​
Indexing Methods

Push MethodProgrammatically upload JSON documents:
  • No restrictions on data source type
  • Real-time updates
  • Full control over connectivity
  • Batch up to 1000 documents or 16 MB
POST /indexes/my-index/docs/index
{
  "value": [
    {
      "@search.action": "upload",
      "id": "1",
      "title": "Azure Search",
      "content": "..."
    }
  ]
}

​
Document Actions

  • upload: Insert new or replace existing documents
  • merge: Update existing document fields
  • mergeOrUpload: Merge if exists, otherwise upload
  • delete: Remove document from index

​
Indexing Workflow

​
AI Enrichment

Add AI-powered transformations during indexing:
  • Data chunking: Split large documents
  • Vectorization: Generate embeddings
  • OCR: Extract text from images
  • Entity recognition: Identify people, places, organizations
  • Key phrase extraction: Extract important terms
  • Language detection: Identify document language

​
Best Practices

Batch Documents

Upload multiple documents per request for efficiency

Handle Errors

Implement retry logic for failed documents

Use Indexers

Automate ingestion from supported sources

Monitor Progress

Track indexer status and document counts

​
Next Steps

Indexers

Learn about automated indexing

Create Index

Build your index schema

Build docs developers (and LLMs) love

Get started for freeTalk to us