gfql cypher pandas: three IN predicates across (a)-[e]->(t) crash with 'Unalignable boolean Series' (polars OK)

[lmeyerov]

Repro (master @ 3fb216d)

import pandas as pd, graphistry
nodes_df = pd.DataFrame({'id': ['a','b','c','tx1','tx2'],'type': ['person','person','company','transaction','transaction']})
edges_df = pd.DataFrame({'src': ['a','b','a','tx1','tx2'],'dst': ['b','c','tx1','tx2','c'],'e_type': ['knows','works_at','sent','transfer','received']})
g = graphistry.edges(edges_df,'src','dst').nodes(nodes_df,'id')
q = "MATCH (a)-[e]->(t) WHERE a.type IN ['person','company'] AND e.e_type IN ['sent','transfer'] AND t.type IN ['transaction','account'] RETURN t.id AS id"
g.gfql(q, engine='polars')   # OK -> ['tx1']
g.gfql(q, engine='pandas')   # GFQLTypeError [invalid-node-reference] Error executing 'rows': Unalignable boolean Series provided as indexer

Replacing the third predicate with t.type = 'transaction' or a map (t {type: 'transaction'}) works on pandas. Engine parity gap: same query, Polars answers, pandas raises. Found while adding tested Cypher twins to gfql/about.rst examples (#2017 follow-up).

🤖 Generated with Claude Code

https://claude.ai/code/session_01WwMmVFo44ADiRRj5cxh7i1

[lmeyerov]

Fix: #2045 (apply single-alias predicate pushdown masks by position) — 77/77 green incl. the polars lane, MERGEABLE on master, awaiting merge; the three-IN shape is pinned there on pandas/polars/cuDF. Closes on merge.