Page History

import System.nanoTime
def profile[R](code: => R, t: Long = nanoTime) = (code, (((nanoTime - t)/1000)).toInt)

import org.apache.hadoop.hbase.io.hfile.CacheConfig
import org.apache.hudi.io.storage.HoodieHFileReader
import org.apache.hadoop.fs.Path
import scala.collection.JavaConverters._


val cacheConfig = new CacheConfig(spark.sparkContext.hadoopConfiguration);
cacheConfig.setCacheDataInL1(false);

val hfilePath = new Path("hdfs://ns-router-dca1/uber-data/tables/temp/satish/hudi-trips-metadata-hfile1tables/my_metadata_table_hfile/default/53dd4e23-012c-4e3e-91fc-9d5ff6a3bf83-0_0-48-91_20210327201656.hfile")
val reader = new HoodieHFileReader(spark.sparkContext.hadoopConfiguration, hfilePath, cacheConfig);

val keys = reader.readAllRecords().asScala.map(x => x.getFirst())
val keyRange = keys.slice(95000, 100000) // pick desired number of keys

var totalTime: Long = 0
var totalRecords: Long = 0
val reader = new HoodieHFileReader(spark.sparkContext.hadoopConfiguration, hfilePath, cacheConfig)
(keyRange).map { k =>
    val (record, time) = profile { reader.getRecordByKey(k) }
    totalTime += time
    if (record.isPresent()) { totalRecords += 1L}
}
"" + (totalTime/1000) + " ms to LOOKUP HFile, #records: " + totalRecords