ai-agents-metadata-filtering-using-multiple-metadatafilters-with-condition-or-o-2bf1c3c6

THEN Upgrade LlamaIndex to a version where this bug is fixed (e.g., after v0.9.44). As a workaround, perform separate queries for each filter condition and merge the results manually, or ensure that the vector store adapter properly supports OR conditions for multiple filters.

# Workaround: query each filter separately and union results
from llama_index.vector_stores.types import MetadataFilters, MetadataFilter

red_filter = MetadataFilters(filters=[MetadataFilter(key="color", value="red")])
alpha_filter = MetadataFilters(filters=[MetadataFilter(key="alpha", value="1")])

red_nodes = index.as_query_engine(filters=red_filter).retrieve(query)
alpha_nodes = index.as_query_engine(filters=alpha_filter).retrieve(query)

# Deduplicate and combine
combined = {node.node_id: node for node in red_nodes + alpha_nodes}
all_nodes = list(combined.values())