gene-bordegaray commented on code in PR #22607: URL: https://github.com/apache/datafusion/pull/22607#discussion_r3323726651
########## datafusion/sqllogictest/test_files/range_partitioning.slt: ########## @@ -0,0 +1,81 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +# The sqllogictest harness registers range_partitioned(range_key, non_range_key, value) +# as an in-memory source with four physical source partitions: +# +# partition 0: range_key in [1, 10), rows (1, 1, 10), (5, 2, 50) +# partition 1: range_key in [10, 20), rows (10, 1, 100), (15, 2, 150) +# partition 2: range_key in [20, 30), rows (20, 1, 200), (25, 2, 250) +# partition 3: range_key in [30, ...), rows (30, 1, 300), (35, 2, 350) + +statement ok +set datafusion.explain.physical_plan_only = true; + +########## +# TEST 1: Aggregate on Range Partition Column +# Scanning range_key preserves source Range partitioning metadata. +# Planning still inserts Hash repartitioning today; later optimizer PRs can +# use this baseline to show when the repartition is removed. +########## + +query TT +EXPLAIN SELECT range_key, SUM(value) FROM range_partitioned GROUP BY range_key; +---- +physical_plan +01)AggregateExec: mode=FinalPartitioned, gby=[range_key@0 as range_key], aggr=[sum(range_partitioned.value)] +02)--RepartitionExec: partitioning=Hash([range_key@0], 4), input_partitions=4 +03)----AggregateExec: mode=Partial, gby=[range_key@0 as range_key], aggr=[sum(range_partitioned.value)] +04)------DataSourceExec: partitions=4, partition_sizes=[1, 1, 1, 1], output_partitioning=Range([range_key@0 ASC], [(10), (20), (30)], 4) + Review Comment: > One thing that bugs me here is that the first and large partitions are infinetely large ([-inf, 10), [10, 20), [20, 30), [30, +inf]). This was part of the design as we decided to always have the whole key space covered (like @asolimando mentioned) > Well, but that problem still exists, nothing prevents you from sending a row to partition [10, 20) when it really belongs to partition [20, 30). This was alaso taken into account with the design of the model. Split points just define each partition's bounds but they do not validate that rows from the source are sent to the correct partition, that is the responsibility of the user declaring the source as range partitioned. We may want to make this more explicit in the docs 👍 -- 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]
