You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@spark.apache.org by Mich Talebzadeh <mi...@gmail.com> on 2020/05/05 15:25:18 UTC

Exception handling in Spark

Hi,

As I understand exception handling in Spark only makes sense if one
attempts an action as opposed to lazy transformations?

Let us assume that I am reading an XML file from the HDFS directory  and
create a dataframe DF on it

val broadcastValue = "123456789"  // I assume this will be sent as a
constant for the batch
// Create a DF on top of XML
val df = spark.read.
                format("com.databricks.spark.xml").
                option("rootTag", "hierarchy").
                option("rowTag", "sms_request").
                load("/tmp/broadcast.xml")

val newDF = df.withColumn("broadcastid", lit(broadcastValue))

newDF.createOrReplaceTempView("tmp")

  // Put data in Hive table
  //
  sqltext = """
  INSERT INTO TABLE michtest.BroadcastStaging PARTITION
(broadcastid="123456", brand)
  SELECT
          ocis_party_id AS partyId
        , target_mobile_no AS phoneNumber
        , brand
        , broadcastid
  FROM tmp
  """
//
// Here I am performing a collection
try  {
         spark.sql(sqltext)
} catch {
    case e: SQLException => e.printStackTrace
    sys.exit()
}

Now the issue I have is that what if the xml file  /tmp/broadcast.xml does
not exist or deleted? I won't be able to catch the error until the hive
table is populated. Of course I can write a shell script to check if the
file exist before running the job or put small collection like
df.show(1,0). Are there more general alternatives?

Thanks

Dr Mich Talebzadeh



LinkedIn * https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw
<https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw>*



http://talebzadehmich.wordpress.com


*Disclaimer:* Use it at your own risk. Any and all responsibility for any
loss, damage or destruction of data or any other property which may arise
from relying on this email's technical content is explicitly disclaimed.
The author will in no case be liable for any monetary damages arising from
such loss, damage or destruction.

Re: Exception handling in Spark

Posted by Brandon Geise <br...@gmail.com>.

Read is an action, so you could wrap it in a Try (or whatever you want)

scala> val df = Try(spark.read.csv("test"))

df: scala.util.Try[org.apache.spark.sql.DataFrame] = Failure(org.apache.spark.sql.AnalysisException: Path does not exist: file:/test;)

From: Mich Talebzadeh <mi...@gmail.com>
Date: Tuesday, May 5, 2020 at 12:45 PM
To: Brandon Geise <br...@gmail.com>
Cc: "user @spark" <us...@spark.apache.org>
Subject: Re: Exception handling in Spark

Thanks  Brandon!

i should have remembered that.

basically the code gets out with sys.exit(1)  if it cannot find the file

I guess there is no easy way of validating DF except actioning it by show(1,0) etc and checking if it works?

Regards,

Dr Mich Talebzadeh

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

http://talebzadehmich.wordpress.com

Disclaimer: Use it at your own risk. Any and all responsibility for any loss, damage or destruction of data or any other property which may arise from relying on this email's technical content is explicitly disclaimed. The author will in no case be liable for any monetary damages arising from such loss, damage or destruction. 

On Tue, 5 May 2020 at 16:41, Brandon Geise <br...@gmail.com> wrote:

You could use the Hadoop API and check if the file exists.

From: Mich Talebzadeh <mi...@gmail.com>
Date: Tuesday, May 5, 2020 at 11:25 AM
To: "user @spark" <us...@spark.apache.org>
Subject: Exception handling in Spark

Hi,

As I understand exception handling in Spark only makes sense if one attempts an action as opposed to lazy transformations?

Let us assume that I am reading an XML file from the HDFS directory  and create a dataframe DF on it

val broadcastValue = "123456789"  // I assume this will be sent as a constant for the batch

// Create a DF on top of XML
val df = spark.read.
                format("com.databricks.spark.xml").
                option("rootTag", "hierarchy").
                option("rowTag", "sms_request").
                load("/tmp/broadcast.xml")

val newDF = df.withColumn("broadcastid", lit(broadcastValue))

newDF.createOrReplaceTempView("tmp")

  // Put data in Hive table
  //
  sqltext = """
  INSERT INTO TABLE michtest.BroadcastStaging PARTITION (broadcastid="123456", brand)
  SELECT
          ocis_party_id AS partyId
        , target_mobile_no AS phoneNumber
        , brand
        , broadcastid
  FROM tmp
  """
//

// Here I am performing a collection 

try  {

         spark.sql(sqltext)

} catch {

    case e: SQLException => e.printStackTrace

    sys.exit()

}

Now the issue I have is that what if the xml file  /tmp/broadcast.xml does not exist or deleted? I won't be able to catch the error until the hive table is populated. Of course I can write a shell script to check if the file exist before running the job or put small collection like df.show(1,0). Are there more general alternatives?

Thanks

Dr Mich Talebzadeh

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

http://talebzadehmich.wordpress.com

Disclaimer: Use it at your own risk. Any and all responsibility for any loss, damage or destruction of data or any other property which may arise from relying on this email's technical content is explicitly disclaimed. The author will in no case be liable for any monetary damages arising from such loss, damage or destruction.

Re: Exception handling in Spark

Posted by Brandon Geise <br...@gmail.com>.

Import scala.util.Try

Import scala.util.Success

Import scala.util.Failure

From: Mich Talebzadeh <mi...@gmail.com>
Date: Tuesday, May 5, 2020 at 6:11 PM
To: Brandon Geise <br...@gmail.com>
Cc: Todd Nist <ts...@gmail.com>, "user @spark" <us...@spark.apache.org>
Subject: Re: Exception handling in Spark

This is what I get

scala> val df = Try(spark.read.format("com.databricks.spark.xml").option("rootTag", "hierarchy").option("rowTag", "sms_request").load("/tmp/broadcast.xml")) Match {case Success(df) => df case Failure(e) => throw new Exception("foo")}
<console>:47: error: not found: value Try
       val df = Try(spark.read.format("com.databricks.spark.xml").option("rootTag", "hierarchy").option("rowTag", "sms_request").load("/tmp/broadcast.xml")) Match {case Success(df) => df case Failure(e) => throw new Exception("foo")}
                ^
<console>:47: error: not found: value Success
       val df = Try(spark.read.format("com.databricks.spark.xml").option("rootTag", "hierarchy").option("rowTag", "sms_request").load("/tmp/broadcast.xml")) Match {case Success(df) => df case Failure(e) => throw new Exception("foo")}
                                                                                                                                                                         ^
<console>:47: error: not found: value Failure
       val df = Try(spark.read.format("com.databricks.spark.xml").option("rootTag", "hierarchy").option("rowTag", "sms_request").load("/tmp/broadcast.xml")) Match {case Success(df) => df case Failure(e) => throw new Exception("foo")}

Dr Mich Talebzadeh

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

http://talebzadehmich.wordpress.com

Disclaimer: Use it at your own risk. Any and all responsibility for any loss, damage or destruction of data or any other property which may arise from relying on this email's technical content is explicitly disclaimed. The author will in no case be liable for any monetary damages arising from such loss, damage or destruction. 

On Tue, 5 May 2020 at 23:03, Brandon Geise <br...@gmail.com> wrote:

This is what I had in mind.  Can you give this approach a try?

val df = Try(spark.read.csv("")) match {

      case Success(df) => df

      case Failure(e) => throw new Exception("foo")

      }

From: Mich Talebzadeh <mi...@gmail.com>
Date: Tuesday, May 5, 2020 at 5:17 PM
To: Todd Nist <ts...@gmail.com>
Cc: Brandon Geise <br...@gmail.com>, "user @spark" <us...@spark.apache.org>
Subject: Re: Exception handling in Spark

I am trying this approach

val broadcastValue = "123456789"  // I assume this will be sent as a constant for the batch
// Create a DF on top of XML
try {
      val df = spark.read.
                format("com.databricks.spark.xml").
                option("rootTag", "hierarchy").
                option("rowTag", "sms_request").
                load("/tmp/broadcast.xml")
          df
}  catch {
    case ex: FileNotFoundException => {
        println (s"\nFile /tmp/broadcast.xml not found\n")
        None
        }
   case unknown: Exception => {
             println(s"\n Error encountered $unknown\n")
        None
        }
}

val newDF = df.withColumn("broadcastid", lit(broadcastValue))

But this does not work

scala> try {
     |       val df = spark.read.
     |                 format("com.databricks.spark.xml").
     |                 option("rootTag", "hierarchy").
     |                 option("rowTag", "sms_request").
     |                 load("/tmp/broadcast.xml")
     |           Some(df)
     | }  catch {
     |     case ex: FileNotFoundException => {
     |         println (s"\nFile /tmp/broadcast.xml not found\n")
     |         None
     |         }
     |    case unknown: Exception => {
     |              println(s"\n Error encountered $unknown\n")
     |         None
     |         }
     | }
res6: Option[org.apache.spark.sql.DataFrame] = Some([brand: string, ocis_party_id: bigint ... 6 more fields])

scala>

scala> df.printSchema
<console>:48: error: not found: value df
       df.printSchema

data frame seems to be lost!

Thanks,

Dr Mich Talebzadeh

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

http://talebzadehmich.wordpress.com

Disclaimer: Use it at your own risk. Any and all responsibility for any loss, damage or destruction of data or any other property which may arise from relying on this email's technical content is explicitly disclaimed. The author will in no case be liable for any monetary damages arising from such loss, damage or destruction. 

On Tue, 5 May 2020 at 18:08, Mich Talebzadeh <mi...@gmail.com> wrote:

Thanks Todd. This is what I did before creating DF on top of that file

var exists = true

exists = xmlDirExists(broadcastStagingConfig.xmlFilePath)

if(!exists) {

  println(s"\n Error: The xml file ${ broadcastStagingConfig.xmlFilePath} does not exist, aborting!\n")

     sys.exit(1)

}

.

.

def xmlFileExists(hdfsDirectory: String): Boolean = {

   val hadoopConf = new org.apache.hadoop.conf.Configuration()

   val fs = org.apache.hadoop.fs.FileSystem.get(hadoopConf)

   fs.exists(new org.apache.hadoop.fs.Path(hdfsDirectory))

 }

And checked it. It works.

Dr Mich Talebzadeh

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

http://talebzadehmich.wordpress.com

Disclaimer: Use it at your own risk. Any and all responsibility for any loss, damage or destruction of data or any other property which may arise from relying on this email's technical content is explicitly disclaimed. The author will in no case be liable for any monetary damages arising from such loss, damage or destruction. 

On Tue, 5 May 2020 at 17:54, Todd Nist <ts...@gmail.com> wrote:

Could you do something like this prior to calling the action.

// Create FileSystem object from Hadoop Configuration

val fs = FileSystem.get(spark.sparkContext.hadoopConfiguration)

// This methods returns Boolean (true - if file exists, false - if file doesn't exist

val fileExists = fs.exists(new Path("<parh_to_file>"))

if (fileExists) println("File exists!")

else println("File doesn't exist!")

Not sure that will help you or not, just a thought.

-Todd

On Tue, May 5, 2020 at 11:45 AM Mich Talebzadeh <mi...@gmail.com> wrote:

Thanks  Brandon!

i should have remembered that.

basically the code gets out with sys.exit(1)  if it cannot find the file

I guess there is no easy way of validating DF except actioning it by show(1,0) etc and checking if it works?

Regards,

Dr Mich Talebzadeh

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

http://talebzadehmich.wordpress.com

Disclaimer: Use it at your own risk. Any and all responsibility for any loss, damage or destruction of data or any other property which may arise from relying on this email's technical content is explicitly disclaimed. The author will in no case be liable for any monetary damages arising from such loss, damage or destruction. 

On Tue, 5 May 2020 at 16:41, Brandon Geise <br...@gmail.com> wrote:

You could use the Hadoop API and check if the file exists.

From: Mich Talebzadeh <mi...@gmail.com>
Date: Tuesday, May 5, 2020 at 11:25 AM
To: "user @spark" <us...@spark.apache.org>
Subject: Exception handling in Spark

Hi,

As I understand exception handling in Spark only makes sense if one attempts an action as opposed to lazy transformations?

Let us assume that I am reading an XML file from the HDFS directory  and create a dataframe DF on it

val broadcastValue = "123456789"  // I assume this will be sent as a constant for the batch

// Create a DF on top of XML
val df = spark.read.
                format("com.databricks.spark.xml").
                option("rootTag", "hierarchy").
                option("rowTag", "sms_request").
                load("/tmp/broadcast.xml")

val newDF = df.withColumn("broadcastid", lit(broadcastValue))

newDF.createOrReplaceTempView("tmp")

  // Put data in Hive table
  //
  sqltext = """
  INSERT INTO TABLE michtest.BroadcastStaging PARTITION (broadcastid="123456", brand)
  SELECT
          ocis_party_id AS partyId
        , target_mobile_no AS phoneNumber
        , brand
        , broadcastid
  FROM tmp
  """
//

// Here I am performing a collection 

try  {

         spark.sql(sqltext)

} catch {

    case e: SQLException => e.printStackTrace

    sys.exit()

}

Now the issue I have is that what if the xml file  /tmp/broadcast.xml does not exist or deleted? I won't be able to catch the error until the hive table is populated. Of course I can write a shell script to check if the file exist before running the job or put small collection like df.show(1,0). Are there more general alternatives?

Thanks

Dr Mich Talebzadeh

LinkedIn  https://www.linkedin.com/profile/view?id=AAEAAAAWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw

http://talebzadehmich.wordpress.com

Disclaimer: Use it at your own risk. Any and all responsibility for any loss, damage or destruction of data or any other property which may arise from relying on this email's technical content is explicitly disclaimed. The author will in no case be liable for any monetary damages arising from such loss, damage or destruction.