Question

My Pinecone vector search stopped returning results after index upgrade due to schema mismatch

Answer 1

Schema mismatches after a Pinecone index upgrade are sneaky because the upsert often succeeds silently but then query returns nothing

import pinecone

pc = Pinecone(api_key="your-api-key")

index = pc.Index("your-index-name")

# First, check what your index actually expects

index_info = pc.describe_index("your-index-name")

print(index_info.dimension)  # what the index expects

# Then check what you're actually sending

import numpy as np

your_vector = np.random.rand(1536)  # Is this matching?

print(len(your_vector))  # must match index dimension exactly

 

Embedding model changed during the upgrade

Super common. If you switched models (even accidentally by upgrading a library), the vector dimensions change:

# OpenAI embedding dimensions by model

# text-embedding-ada-002  → 1536 dimensions

# text-embedding-3-small  → 1536 (default) or custom

# text-embedding-3-large  → 3072 dimensions

# So if you upserted with ada-002 but now query with 3-large

# you'll get nothing back — dimension mismatch

from openai import OpenAI

client = OpenAI()

def get_embedding(text, model="text-embedding-ada-002"):

    response = client.embeddings.create(input=text, model=model)

    print(f"Embedding dimension: {len(response.data[0].embedding)}")

    return response.data[0].embedding