andygrove opened a new issue, #4560: URL: https://github.com/apache/datafusion-comet/issues/4560
## Description `array_size(NULL)` should return `NULL`, but Comet returns `-1`. `array_size` is `RuntimeReplaceable` and lowers to `Size(child, legacySizeOfNull = false)`. With `legacySizeOfNull = false`, Spark returns `NULL` for a NULL input; with `legacySizeOfNull = true` it returns `-1`. Comet's native `Size` implementation appears to always return `-1` for NULL, ignoring the `legacySizeOfNull = false` case. Found while sweeping `RuntimeReplaceable` functions for Comet support (the query runs natively via `CometProject`, so it does not fall back; it just returns the wrong value). ## Likely broader impact The same `legacySizeOfNull = false` path is used by `size` and `cardinality` (whose flag comes from `spark.sql.legacy.sizeOfNull`, default `false` in modern Spark), so they may exhibit the same NULL-handling bug. Worth verifying. ## Repro (SQL) ```sql SELECT array_size(CAST(NULL AS ARRAY<INT>)); -- Spark: NULL, Comet: -1 ``` ## Expected Match Spark: return `NULL` when `legacySizeOfNull = false` and the input is NULL. -- 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]
