alexeykudinkin commented on a change in pull request #4948:
URL: https://github.com/apache/hudi/pull/4948#discussion_r822043839
##########
File path:
hudi-spark-datasource/hudi-spark/src/test/scala/org/apache/hudi/TestHoodieFileIndex.scala
##########
@@ -333,6 +333,57 @@ class TestHoodieFileIndex extends HoodieClientTestBase {
assert(fileIndex.getAllQueryPartitionPaths.get(0).path.equals("c"))
}
+ @Test
+ def testDataSkippingWhileFileListing(): Unit = {
+ val r = new Random(0xDEED)
+ val tuples = for (i <- 1 to 1000) yield (i, 1000 - i, r.nextString(5),
r.nextInt(4))
+
+ val _spark = spark
+ import _spark.implicits._
+ val inputDF = tuples.toDF("id", "inv_id", "str", "rand")
+
+ val opts = Map(
+ "hoodie.insert.shuffle.parallelism" -> "4",
+ "hoodie.upsert.shuffle.parallelism" -> "4",
+ HoodieWriteConfig.TBL_NAME.key -> "hoodie_test",
+ RECORDKEY_FIELD.key -> "id",
+ PRECOMBINE_FIELD.key -> "id",
+ HoodieMetadataConfig.ENABLE.key -> "true",
+ HoodieMetadataConfig.ENABLE_METADATA_INDEX_COLUMN_STATS.key -> "true",
+
HoodieMetadataConfig.ENABLE_METADATA_INDEX_COLUMN_STATS_FOR_ALL_COLUMNS.key ->
"true",
+ HoodieTableConfig.POPULATE_META_FIELDS.key -> "true"
+ )
+
+ inputDF.repartition(4)
+ .write
+ .format("hudi")
+ .options(opts)
+ .option(DataSourceWriteOptions.OPERATION.key,
DataSourceWriteOptions.INSERT_OPERATION_OPT_VAL)
+ .option(HoodieStorageConfig.PARQUET_MAX_FILE_SIZE.key, 100 * 1024)
+ .mode(SaveMode.Overwrite)
+ .save(basePath)
+
+ metaClient = HoodieTableMetaClient.reload(metaClient)
+
+ val props = Map[String, String](
+ "path" -> basePath,
+ QUERY_TYPE.key -> QUERY_TYPE_SNAPSHOT_OPT_VAL,
+ DataSourceReadOptions.ENABLE_DATA_SKIPPING.key -> "true"
+ )
+
+ val fileIndex = HoodieFileIndex(spark, metaClient, Option.empty, props,
NoopCache)
+
+ val allFilesPartitions = fileIndex.listFiles(Seq(), Seq())
+ assertEquals(10, allFilesPartitions.head.files.length)
+
+ // We're selecting a single file that contains "id" == 1 row, which there
should be
+ // strictly 1. Given that 1 is minimal possible value, Data Skipping
should be able to
+ // truncate search space to just a single file
+ val dataFilter = EqualTo(AttributeReference("id", IntegerType, nullable =
false)(), Literal(1))
Review comment:
This is tested separately in unit-tests `TestDataSkippingUtils`
--
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]
