Updated on 2022-12-14 GMT+08:00

Reading COW Table Views

  • Reading the real-time view (using Hive and SparkSQL as an example): Directly read the Hudi table stored in Hive.
    select count(*) from test;
  • Reading the real-time view (using the Spark DataSource API as an example): This is similar to reading a common DataSource table.

    QUERY_TYPE_OPT_KEY must be set to QUERY_TYPE_SNAPSHOT_OPT_VAL.

    spark.read.format("hudi")
    .option(QUERY_TYPE_OPT_KEY, QUERY_TYPE_SNAPSHOT_OPT_VAL) // Set the query type to the real-time view.
    .load("/tmp/default/cow_bugx/*/*/*/*") // Set the path of the Hudi table to be read. The current table has three levels of partitions.
    .createTempView("mycall")
    spark.sql("select * from mycall").show(100)
  • Reading the incremental view (using Hive and SparkSQL as an example):
    set hoodie.test.consume.mode=INCREMENTAL;  // Specify the incremental reading mode.
    set hoodie.test.consume.max.commits=3;  // Specify the maximum number of commits to be consumed.
    set hoodie.test.consume.start.timestamp=20201227153030;  // Specify the initial incremental pull commit.
    select count(*) from default.test where `_hoodie_commit_time`>'20201227153030'; // This filtering condition must be added, and the value is the initial incremental pull commit.
  • Reading the incremental view (using the Spark DataSource API as an example):

    QUERY_TYPE_OPT_KEY must be set to QUERY_TYPE_INCREMENTAL_OPT_VAL.

    spark.read.format("hudi")  
    .option(QUERY_TYPE_OPT_KEY, QUERY_TYPE_INCREMENTAL_OPT_VAL) // Set the query type to the incremental mode.
    .option(BEGIN_INSTANTTIME_OPT_KEY, "20210308212004")  // Specify the initial incremental pull commit.
    .option(END_INSTANTTIME_OPT_KEY, "20210308212318")  //: Specify the end commit of the incremental pull.
    .load("/tmp/default/cow_bugx/*/*/*/*")  // Set the path of the Hudi table to be read. The current table has three levels of partitions.
    .createTempView("mycall")  // Register as a Spark temporary table.
    spark.sql("select * from mycall where `_hoodie_commit_time`>'20210308211131'")// Start the query. The statement is the same as the Hive incremental query statement.
    .show(100, false)
  • Reading the read-optimized view: The read-optimized view of COW tables is equivalent to the real-time view.