Rich-T-kid commented on code in PR #24227: URL: https://github.com/apache/datafusion/pull/24227#discussion_r3962343162
########## datafusion/sqllogictest/test_files/parquet_rle_to_dictionary.slt: ########## @@ -0,0 +1,334 @@ +# 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. + +# Tests for datafusion.execution.parquet.enable_rle_to_dictionary. The flag +# applies only when DataFusion infers the table schema; explicit schemas are +# not promoted. + +# Write with dictionary_enabled=true to guarantee RLE_DICTIONARY encoding. +query I +COPY ( + SELECT column1 AS product_category, column2 AS status + FROM (VALUES + ('electronics', 'active'), + ('clothing', 'active'), + ('electronics', 'inactive'), + ('books', 'inactive'), + ('furniture', 'pending'), + ('books', 'active') + ) +) +TO 'test_files/scratch/parquet_rle_to_dictionary/products.parquet' +STORED AS PARQUET +OPTIONS ('format.dictionary_enabled' true); +---- +6 + +statement ok +set datafusion.execution.parquet.enable_rle_to_dictionary = false; + +statement ok +CREATE EXTERNAL TABLE products_utf8 +STORED AS PARQUET +LOCATION 'test_files/scratch/parquet_rle_to_dictionary/products.parquet'; + +query TT +SELECT DISTINCT arrow_typeof(product_category), arrow_typeof(status) FROM products_utf8; +---- +Utf8View Utf8View + +query TI rowsort +SELECT product_category, COUNT(*) FROM products_utf8 GROUP BY product_category; +---- +books 2 +clothing 1 +electronics 2 +furniture 1 + +statement ok +DROP TABLE products_utf8; + +statement ok +set datafusion.execution.parquet.enable_rle_to_dictionary = true; + +statement ok +CREATE EXTERNAL TABLE products_dict +STORED AS PARQUET +LOCATION 'test_files/scratch/parquet_rle_to_dictionary/products.parquet'; + +query TT +SELECT DISTINCT arrow_typeof(product_category), arrow_typeof(status) FROM products_dict; +---- +Dictionary(Int32, Utf8) Dictionary(Int32, Utf8) + +query TI rowsort +SELECT product_category, COUNT(*) FROM products_dict GROUP BY product_category; +---- +books 2 +clothing 1 +electronics 2 +furniture 1 + +# Predicates and aggregates must work against dictionary scan output. +query TT rowsort +SELECT product_category, status FROM products_dict WHERE status = 'active'; +---- +books active +clothing active +electronics active + +query TI rowsort +SELECT product_category, COUNT(*) FROM products_dict WHERE status != 'inactive' GROUP BY product_category; +---- +books 1 +clothing 1 +electronics 1 +furniture 1 + +query TT +SELECT DISTINCT arrow_typeof(product_category), arrow_typeof(product_count) +FROM ( + SELECT product_category, COUNT(*) AS product_count + FROM products_dict + WHERE status != 'inactive' + GROUP BY product_category +); +---- +Dictionary(Int32, Utf8) Int64 + +statement ok +DROP TABLE products_dict; + +# Per-column writer options: only columns with dictionary pages are promoted. + +statement ok +COPY ( + SELECT column1 AS env, column2 AS region, column3 AS build + FROM (VALUES + ('prod', 'us-east', 'debug'), + ('staging', 'us-west', 'release'), + ('prod', 'us-east', 'debug') + ) +) +TO 'test_files/scratch/parquet_rle_to_dictionary/selective.parquet' +STORED AS PARQUET +OPTIONS ( + 'format.dictionary_enabled' false, + 'format.dictionary_enabled::env' true, + 'format.dictionary_enabled::region' true +); + +statement ok +set datafusion.execution.parquet.enable_rle_to_dictionary = false; + +statement ok +CREATE EXTERNAL TABLE selective_utf8 +STORED AS PARQUET +LOCATION 'test_files/scratch/parquet_rle_to_dictionary/selective.parquet'; + +query TTT +SELECT DISTINCT arrow_typeof(env), arrow_typeof(region), arrow_typeof(build) +FROM selective_utf8; +---- +Utf8View Utf8View Utf8View + +statement ok +DROP TABLE selective_utf8; + +statement ok +set datafusion.execution.parquet.enable_rle_to_dictionary = true; + +statement ok +CREATE EXTERNAL TABLE selective_dict +STORED AS PARQUET +LOCATION 'test_files/scratch/parquet_rle_to_dictionary/selective.parquet'; + +query TTT +SELECT DISTINCT arrow_typeof(env), arrow_typeof(region), arrow_typeof(build) +FROM selective_dict; +---- +Dictionary(Int32, Utf8) Dictionary(Int32, Utf8) Utf8View + +statement ok +DROP TABLE selective_dict; + +# Cross-file mixed encoding: one file has dictionary pages and one is plain. +# Schema inference must merge both as Dictionary(Int32, Utf8). + +statement ok +COPY (SELECT column1 AS category FROM (VALUES ('electronics'), ('electronics'))) +TO 'test_files/scratch/parquet_rle_to_dictionary/mixed/rle.parquet' +STORED AS PARQUET OPTIONS ('format.dictionary_enabled' true); + +statement ok +COPY (SELECT column1 AS category FROM (VALUES ('books'), ('books'))) +TO 'test_files/scratch/parquet_rle_to_dictionary/mixed/plain.parquet' +STORED AS PARQUET OPTIONS ('format.dictionary_enabled' false); + +statement ok +set datafusion.execution.parquet.enable_rle_to_dictionary = true; + +statement ok +CREATE EXTERNAL TABLE mixed_encoding +STORED AS PARQUET +LOCATION 'test_files/scratch/parquet_rle_to_dictionary/mixed/'; + +query TI +SELECT arrow_typeof(category), COUNT(*) FROM mixed_encoding GROUP BY arrow_typeof(category); +---- +Dictionary(Int32, Utf8) 4 + +statement ok +DROP TABLE mixed_encoding; + +# Mixed encoding with high cardinality: the plain file contains more than 128 +# distinct values so a narrow (Int8) dictionary key would overflow. After the +# fix, uniform_dict_schemas widens the shared key to at least Int32. +# Both files must be scannable and return the correct row count. + +statement ok +COPY ( + SELECT v AS category + FROM ( + VALUES + ('v000'),('v001'),('v002'),('v003'),('v004'),('v005'),('v006'),('v007'), + ('v008'),('v009'),('v010'),('v011'),('v012'),('v013'),('v014'),('v015'), + ('v016'),('v017'),('v018'),('v019'),('v020'),('v021'),('v022'),('v023'), + ('v024'),('v025'),('v026'),('v027'),('v028'),('v029'),('v030'),('v031'), + ('v032'),('v033'),('v034'),('v035'),('v036'),('v037'),('v038'),('v039'), + ('v040'),('v041'),('v042'),('v043'),('v044'),('v045'),('v046'),('v047'), + ('v048'),('v049'),('v050'),('v051'),('v052'),('v053'),('v054'),('v055'), + ('v056'),('v057'),('v058'),('v059'),('v060'),('v061'),('v062'),('v063'), + ('v064'),('v065'),('v066'),('v067'),('v068'),('v069'),('v070'),('v071'), + ('v072'),('v073'),('v074'),('v075'),('v076'),('v077'),('v078'),('v079'), + ('v080'),('v081'),('v082'),('v083'),('v084'),('v085'),('v086'),('v087'), + ('v088'),('v089'),('v090'),('v091'),('v092'),('v093'),('v094'),('v095'), + ('v096'),('v097'),('v098'),('v099'),('v100'),('v101'),('v102'),('v103'), + ('v104'),('v105'),('v106'),('v107'),('v108'),('v109'),('v110'),('v111'), + ('v112'),('v113'),('v114'),('v115'),('v116'),('v117'),('v118'),('v119'), + ('v120'),('v121'),('v122'),('v123'),('v124'),('v125'),('v126'),('v127'), + ('v128'),('v129') + ) AS t(v) +) +TO 'test_files/scratch/parquet_rle_to_dictionary/high_card/plain.parquet' +STORED AS PARQUET OPTIONS ('format.dictionary_enabled' false); + +statement ok +COPY (SELECT column1 AS category FROM (VALUES ('rle_a'), ('rle_b'), ('rle_a'))) +TO 'test_files/scratch/parquet_rle_to_dictionary/high_card/rle.parquet' +STORED AS PARQUET OPTIONS ('format.dictionary_enabled' true); Review Comment: updated the test to cast to a int8 first before overflowing https://github.com/apache/datafusion/commit/4d607b74af918f514fb4f521b9b307a3458a5546 -- 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]
