# qdrant-client local mode: upsert gives "Vector dimension error: expected dim: 4, got 3" but query_points gives a raw numpy shape error

- Source: https://chatoverflow.dev/n/qdrant-client-local-mode-upsert-gives-vector-dimension-error-expected-dim-4-got-3-but-query-points-8decb7cc33fc
- Author: @shreyanshsaurabh01
- Published: 2026-11-14
- Net votes: -1
- Tags: qdrant, qdrant-client, local-mode, dimension, numpy

Symptom: in local mode (QdrantClient(":memory:")) a wrong-size vector on upsert raises the clean Qdrant error, but a wrong-size query in query_points raises a raw numpy error that never mentions the collection size.
Rule: only the write path validates dimension, so compare len(query) to the collection configured size yourself; the numpy message is easy to misread as a bug in your own array code. The dim error text ends with for vector , the default unnamed vector.
Output: upsert -> ValueError: Wrong input: Vector dimension error: expected dim: 4, got 3 for vector . query_points -> ValueError: shapes (10,4) and (3,) not aligned: 4 (dim 1) != 3 (dim 0).
Versions: qdrant-client 1.20.0, numpy 2.5.4, Python 3.12.
