[
https://issues.apache.org/jira/browse/SPARK-40637?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
xsys updated SPARK-40637:
-------------------------
Summary: Spark-shell can correctly encode BINARY type but Spark-sql cannot
(was: DataFrame can correctly encode BINARY type but SparkSQL cannot)
> Spark-shell can correctly encode BINARY type but Spark-sql cannot
> -----------------------------------------------------------------
>
> Key: SPARK-40637
> URL: https://issues.apache.org/jira/browse/SPARK-40637
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 3.2.1
> Reporter: xsys
> Priority: Major
> Attachments: image-2022-10-18-12-15-05-576.png
>
>
> h3. Describe the bug
> When we store a BINARY value (e.g. {{BigInt("1").toByteArray)}} /
> {{{}X'01'{}}}) either via {{spark-shell or spark-sql, and then read it from
> Spark-shell, it}} outputs {{{}[01]{}}}. However, it does not encode correctly
> via {{{}spark-sql{}}}.
> h3. To Reproduce
> On Spark 3.2.1 (commit {{{}4f25b3f712{}}}), using {{{}spark-shell{}}}:
> {code:java}
> $SPARK_HOME/bin/spark-shell{code}
> Execute the following:
> {code:java}
> scala> val rdd = sc.parallelize(Seq(Row(BigInt("1").toByteArray)))
> rdd: org.apache.spark.rdd.RDD[org.apache.spark.sql.Row] =
> ParallelCollectionRDD[356] at parallelize at <console>:28
> scala> val schema = new StructType().add(StructField("c1", BinaryType, true))
> schema: org.apache.spark.sql.types.StructType =
> StructType(StructField(c1,BinaryType,true))
> scala> val df = spark.createDataFrame(rdd, schema)
> df: org.apache.spark.sql.DataFrame = [c1: binary]
> scala> df.show(false)
> +----+
> |c1 |
> +----+
> |[01]|
> +----+
> scala> df.write.mode("overwrite").format("orc").saveAsTable("binary_vals")
> scala> spark.sql("select * from binary_vals;").show(false)
> +----+
> |c1 |
> +----+
> |[01]|
> +----+{code}
> We can see the output using "spark.sql" in spark-shell.
> Then using {{spark-sql}} (to see we use tee to redirect the log to a file)
> {code:java}
> $SPARK_HOME/bin/spark-sql | tee sql.log{code}
> Execute the following, we only get an empty output in the terminal:
> {code:java}
> spark-sql> select * from binary_vals; -- check what was inserted by DataFrame
> spark-sql> drop table binary_vals;
> spark-sql> create table binary_vals(c1 BINARY) stored as ORC; -- try to
> insert directly in spark-sql
> spark-sql> insert into binary_vals select X'01';
> spark-sql> select * from binary_vals;
> Time taken: 0.077 seconds, Fetched 1 row(s)
> {code}
>
> From the log file, we find it shows as a garbage character. (We never
> encountered this garbage character in logs of other data types)
> h3. !image-2022-10-18-12-15-05-576.png!
> h3. Expected behavior
> We expect the two Spark interfaces ({{{}spark-sql{}}} & {{{}spark-shell{}}})
> to behave consistently for the same data type ({{{}BINARY{}}}) & input
> ({{{}BigInt("1").toByteArray){}}} / {{{}X'01'{}}}) combination.
>
> h3. Additional context
> We also tried Avro and Parquet and encountered the same issue. We believe
> this is format-independent.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
