Hudi Clustering

0

I am using EMR 6.6.0, which has hudi 10.1. I am trying to bulkinsert and do inline clustering using Hudi. But seems its not clustering the file as per file size being mentioned. But it is still producing the files in KB only.

I tried below configuration:

hudi_clusteringopt = { 'hoodie.table.name': 'myhudidataset_upsert_legacy_new7', 'hoodie.datasource.write.recordkey.field': 'id', 'hoodie.datasource.write.partitionpath.field': 'creation_date', 'hoodie.datasource.write.precombine.field': 'last_update_time', 'hoodie.datasource.hive_sync.enable': 'true', 'hoodie.datasource.hive_sync.database': 'my_hudi_db', 'hoodie.datasource.hive_sync.table': 'myhudidataset_upsert_legacy_new7', 'hoodie.datasource.hive_sync.partition_fields': 'creation_date', 'hoodie.datasource.hive_sync.partition_extractor_class': 'org.apache.hudi.hive.MultiPartKeysValueExtractor', "hoodie.datasource.write.hive_style_partitioning": "true", "hoodie.datasource.write.operation": "bulk_insert", }

"hoodie.datasource.write.operation": "bulk_insert",

try:

inputDF.write.format("org.apache.hudi"). \
        options(**hudi_clusteringopt). \
        option("hoodie.parquet.small.file.limit", "0"). \
        option("hoodie.clustering.inline", "true"). \
        option("hoodie.clustering.inline.max.commits", "0"). \
        option("hoodie.clustering.plan.strategy.target.file.max.bytes", "1073741824"). \
        option("hoodie.clustering.plan.strategy.small.file.limit", "629145600"). \
        option("hoodie.clustering.plan.strategy.sort.columns", "pk_col"). \
        mode('append'). \
        save("s3://xxxxxxxxxxxxxx");

except Exception as e: print(e)

Here is the data set if someone wants to regenerate:

inputDF = spark.createDataFrame( [ ("1001",1001, "2015-01-01", "2015-01-01T13:51:39.340396Z"), ("1011",1011, "2015-01-01", "2015-01-01T12:14:58.597216Z"), ("1021",1021, "2015-01-01", "2015-01-01T13:51:40.417052Z"), ("1031",1031, "2015-01-01", "2015-01-01T13:51:40.519832Z"), ("1041",1041, "2015-01-02", "2015-01-01T12:15:00.512679Z"), ("1051",1051, "2015-01-02", "2015-01-01T13:51:42.248818Z"), ], ["id","id_val", "creation_date", "last_update_time"] )

AWS
Zahid
已提问 2 年前409 查看次数
1 回答
0

Not certain, but I would give this parameter a try next:
"hoodie.copyonwrite.record.size.estimate" (https://hudi.apache.org/docs/configurations/#hoodiecopyonwriterecordsizeestimate)

This is the parameter I use to assist with file sizing. The lower you set it the larger the files. The documentation mentions this should be automatically determined but I've always had to manually tune this parameter to avoid small files.

tjtoll
已回答 2 年前
  • I tried with EMR but its not updating the file size

您未登录。 登录 发布回答。

一个好的回答可以清楚地解答问题和提供建设性反馈,并能促进提问者的职业发展。

回答问题的准则