Considerations when using the Parquet storage type in the Spark1.2.0 version:
SQL statements:
Select * from Order_created_dynamic_partition_parquet;
To perform the results in Spark-sql:
the - to [email protected] [[email protected]2014-05 [Email protected] [ [email protected]2014-05 [] [ email protected] [[email protected]2014-05 [email protected] [email protected]2014-05 [Email protected]< C12/>[[email protected]
To perform the results in Beeline:
Error:
[B cannot is cast to java.lang.String (state=,code=0)
To perform the results in hive:
ordernumber event_time event_month10703007267488 the- to- on .: on:12.334+ on the- to10101043505096 the- to- on -: -:12.342+ on the- to10103043509747 the- to- on -: -:12.33+ on the- to10103043501575 the- to- on the: -:12.33+ on the- to10104043514061 the- to- on the:Geneva:12.324+ on the- to
Can be set by setting
set spark.sql.parquet.binaryAsString=true
To resolve problems in Spark-sql and Beeline, the default value of this parameter is false in the spark1.2.0 version;
Description: Some Other parquet-producing systems, in particular Impala and older versions of Spark SQL, does not differentiate Betwee n binary data and strings when writing out the parquet schema. This flag tells Spark SQL to interpret binary data as a string to provide compatibility with these systems.
spark1.2.0 version Sparksql using parquet type considerations