ai-agents-metadata-filter-or-multiple-metadatafilters-with-or-condition-on-qdra-ab8ade8c

THEN Perform separate queries for each filter and merge the results (e.g., union of node sets). Alternatively, upgrade to a newer version of LlamaIndex that correctly processes OR conditions on Qdrant, or verify that the vector store supports OR logic.

from llama_index import VectorStoreIndex

# Workaround: query separately and merge
query_engine_red = index.as_query_engine(filters=MetadataFilters(filters=[MetadataFilter(key="color", value="red")]))
query_engine_alpha = index.as_query_engine(filters=MetadataFilters(filters=[MetadataFilter(key="alpha", value="1")]))
results_red = query_engine_red.query("query text")
results_alpha = query_engine_alpha.query("query text")
# Merge nodes from results_red and results_alpha