parthchandra commented on PR #6413: URL: https://github.com/apache/datafusion-comet/pull/6413#issuecomment-5919287580
Two coverage notes (Spark canonicalizes NaN and folds `-0.0` before hashing): - **`spark/src/test/resources/sql-tests/expressions/hash/hash.sql:47`** and the `shapes` closure at **`native/spark-expr/src/hash_funcs/murmur3.rs:833`** — the nested coverage tests arrays, structs, dictionaries, and fixed-size/large lists, but not maps. Map float keys and values do go through `hash_input` via the recursive fallback, so it should already be correct, but nothing tests it. Please add a map-of-float case in both. - **`native/spark-expr/src/agg_funcs/hll_plus_plus.rs:507`** — the `floats_fold_negative_zero_and_nan` test only exercises top-level Float64/Float32. Dropping the old `normalize_floats` pass actually improves nested floats (that pass never reached a float inside a struct or array, while the hash kernel does). Please add an `approx_count_distinct` case over a struct-of-float or array-of-float so that improvement can't regress. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
