lanyuanxiaoyao/hudi
1
0
Fork 0
Code Activity

[HUDI-704] Add test for RepairsCommand (#1554)

Browse Source
This commit is contained in:
hongdd
2020-05-07 23:02:28 +08:00
committed by GitHub
parent 366bb10d8c
commit f921469afc
7 changed files with 457 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
hudi-cli/src/main/scala/org/apache/hudi/cli/SparkHelpers.scala
View File

@@ -46,6 +46,10 @@ object SparkHelpers {
HoodieIndexConfig.DEFAULT_HOODIE_BLOOM_INDEX_FILTER_DYNAMIC_MAX_ENTRIES.toInt, HoodieIndexConfig.DEFAULT_BLOOM_INDEX_FILTER_TYPE);
val writeSupport: HoodieAvroWriteSupport = new HoodieAvroWriteSupport(new AvroSchemaConverter().convert(schema), schema, filter)
val parquetConfig: HoodieParquetConfig = new HoodieParquetConfig(writeSupport, CompressionCodecName.GZIP, HoodieStorageConfig.DEFAULT_PARQUET_BLOCK_SIZE_BYTES.toInt, HoodieStorageConfig.DEFAULT_PARQUET_PAGE_SIZE_BYTES.toInt, HoodieStorageConfig.DEFAULT_PARQUET_FILE_MAX_BYTES.toInt, fs.getConf, HoodieStorageConfig.DEFAULT_STREAM_COMPRESSION_RATIO.toDouble)
// Add current classLoad for config, if not will throw classNotFound of 'HoodieWrapperFileSystem'.
parquetConfig.getHadoopConf().setClassLoader(Thread.currentThread.getContextClassLoader)
val writer = new HoodieParquetWriter[HoodieJsonPayload, IndexedRecord](instantTime, destinationFile, parquetConfig, schema, new SparkTaskContextSupplier())
for (rec <- sourceRecords) {
val key: String = rec.get(HoodieRecord.RECORD_KEY_METADATA_FIELD).toString