pgvector "different vector dimensions" Error

How to fix it

// try these in order, the most common fix is first

1. FIX 1Use one model per column

   Copy steps
   1. 01Check the column with \d table
   2. 02Confirm your model's output size
   Did this fix it? Yes No

2. FIX 2Add a column or table

   Copy steps
   1. 01ALTER TABLE ADD COLUMN embedding2 vector(768)
   2. 02Re-embed rows and swap queries
   Did this fix it? Yes No

3. FIX 3Re-embed everything

   Copy steps
   1. 01Recreate the table at the new dimension
   2. 02Rebuild the index after loading
   Did this fix it? Yes No

Symptoms

ERROR: expected 1536 dimensions, not 768, or different vector dimensions 1536 and 768.

Why it happens

• C-01Changed embedding model
• C-02Tool created the table from the first model
• C-03Mixed models in one table

// this guide also covers

• different vector dimensions 1536 and 768
• expected 1536 dimensions not 768
• pgvector dimension mismatch

The same problem in other apps

All not loading, error messages & connection guides →

• Pinecone 429 Too Many Requests Error (Pinecone)
• Weaviate Error: Vector Lengths Don't Match (Weaviate)
• Chroma "Your system has an unsupported version of sqlite3" Error (Chroma)
• Milvus "collection not loaded" Error (Milvus)
• Qdrant "Wrong input: Vector dimension error: expected dim, got" (Qdrant)

Other pgvector errors

• pgvector Error: extension "vector" is not available
• pgvector HNSW Index Build Slow
• pgvector "operator does not exist: vector <=>" Error
• pgvector Dimension Limit: Cannot Index More Than 2000 Dimensions
• pgvector Query Returns Fewer Results Than LIMIT With Filter
• pgvector Index Not Being Used
• pgvector HNSW Build Out of Memory or Shared Memory Error
• How to Enable pgvector on Supabase