Major Description
In Hive tests, in the Scala 2.12 build, still seeing a few failures that seem to show mismatching schema inference. Not clear whether it's the same as SPARK-25044. Examples:
- SPARK-5775 read array from partitioned_parquet_with_key_and_complextypes *** FAILED *** Results do not match for query: Timezone: sun.util.calendar.ZoneInfo[id="America/Los_Angeles",offset=-28800000,dstSavings=3600000,useDaylight=true,transitions=185,lastRule=java.util.SimpleTimeZone[id=America/Los_Angeles,offset=-28800000,dstSavings=3600000,useDaylight=true,startYear=0,startMode=3,startMonth=2,startDay=8,startDayOfWeek=1,startTime=7200000,startTimeMode=0,endMode=3,endMonth=10,endDay=1,endDayOfWeek=1,endTime=7200000,endTimeMode=0]] Timezone Env: == Parsed Logical Plan == 'Project ['arrayField, 'p] +- 'Filter ('p = 1) +- 'UnresolvedRelation `partitioned_parquet_with_key_and_complextypes` == Analyzed Logical Plan == arrayField: array<int>, p: int Project [arrayField#82569, p#82570] +- Filter (p#82570 = 1) +- SubqueryAlias `default`.`partitioned_parquet_with_key_and_complextypes` +- Relation[intField#82566,stringField#82567,structField#82568,arrayField#82569,p#82570] parquet == Optimized Logical Plan == Project [arrayField#82569, p#82570] +- Filter (isnotnull(p#82570) && (p#82570 = 1)) +- Relation[intField#82566,stringField#82567,structField#82568,arrayField#82569,p#82570] parquet == Physical Plan == *(1) Project [arrayField#82569, p#82570] +- *(1) FileScan parquet default.partitioned_parquet_with_key_and_complextypes[arrayField#82569,p#82570] Batched: false, Format: Parquet, Location: PrunedInMemoryFileIndex[file:/home/srowen/spark-2.12/sql/hive/target/tmp/spark-d8d87d74-33e7-4f22..., PartitionCount: 1, PartitionFilters: [isnotnull(p#82570), (p#82570 = 1)], PushedFilters: [], ReadSchema: struct<arrayField:array<int>> == Results == == Results == !== Correct Answer - 10 == == Spark Answer - 10 == !struct<> struct<arrayField:array<int>,p:int> ![Range 1 to 1,1] [WrappedArray(1),1] ![Range 1 to 10,1] [WrappedArray(1, 2),1] ![Range 1 to 2,1] [WrappedArray(1, 2, 3),1] ![Range 1 to 3,1] [WrappedArray(1, 2, 3, 4),1] ![Range 1 to 4,1] [WrappedArray(1, 2, 3, 4, 5),1] ![Range 1 to 5,1] [WrappedArray(1, 2, 3, 4, 5, 6),1] ![Range 1 to 6,1] [WrappedArray(1, 2, 3, 4, 5, 6, 7),1] ![Range 1 to 7,1] [WrappedArray(1, 2, 3, 4, 5, 6, 7, 8),1] ![Range 1 to 8,1] [WrappedArray(1, 2, 3, 4, 5, 6, 7, 8, 9),1] ![Range 1 to 9,1] [WrappedArray(1, 2, 3, 4, 5, 6, 7, 8, 9, 10),1] (QueryTest.scala:163)
- SPARK-2693 udaf aggregates test *** FAILED *** Results do not match for query: Timezone: sun.util.calendar.ZoneInfo[id="America/Los_Angeles",offset=-28800000,dstSavings=3600000,useDaylight=true,transitions=185,lastRule=java.util.SimpleTimeZone[id=America/Los_Angeles,offset=-28800000,dstSavings=3600000,useDaylight=true,startYear=0,startMode=3,startMonth=2,startDay=8,startDayOfWeek=1,startTime=7200000,startTimeMode=0,endMode=3,endMonth=10,endDay=1,endDayOfWeek=1,endTime=7200000,endTimeMode=0]] Timezone Env: == Parsed Logical Plan == 'GlobalLimit 1 +- 'LocalLimit 1 +- 'Project [unresolvedalias('percentile('key, 'array(1, 1)), None)] +- 'UnresolvedRelation `src` == Analyzed Logical Plan == percentile(key, array(1, 1), 1): array<double> GlobalLimit 1 +- LocalLimit 1 +- Aggregate [percentile(key#205098, cast(array(1, 1) as array<double>), 1, 0, 0) AS percentile(key, array(1, 1), 1)#205101] +- SubqueryAlias `default`.`src` +- HiveTableRelation `default`.`src`, org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, [key#205098, value#205099] == Optimized Logical Plan == GlobalLimit 1 +- LocalLimit 1 +- Aggregate [percentile(key#205098, [1.0,1.0], 1, 0, 0) AS percentile(key, array(1, 1), 1)#205101] +- Project [key#205098] +- HiveTableRelation `default`.`src`, org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, [key#205098, value#205099] == Physical Plan == CollectLimit 1 +- ObjectHashAggregate(keys=[], functions=[percentile(key#205098, [1.0,1.0], 1, 0, 0)], output=[percentile(key, array(1, 1), 1)#205101]) +- Exchange SinglePartition +- ObjectHashAggregate(keys=[], functions=[partial_percentile(key#205098, [1.0,1.0], 1, 0, 0)], output=[buf#205104]) +- Scan hive default.src [key#205098], HiveTableRelation `default`.`src`, org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, [key#205098, value#205099] == Results == == Results == !== Correct Answer - 1 == == Spark Answer - 1 == !struct<array(max(key), max(key)):array<int>> struct<percentile(key, array(1, 1), 1):array<double>> ![WrappedArray(498, 498)] [WrappedArray(498.0, 498.0)] (QueryTest.scala:163)
