[
https://issues.apache.org/jira/browse/SPARK-34529?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17405192#comment-17405192
]
Syedhamjath commented on SPARK-34529:
-------------------------------------
I'm having similar issue as [~kc.shanmugavel] If we go by documentation \r\n is
not a single character it is two characters. Therefore, lineSep option should
accept more than one character.
To understand it better I have created a test data - [^TestData.csv]
This TestData.csv is UTF-16 BE BOM and CRLF is used as record delimiter.
!image-2021-08-26-14-06-41-055.png!
Scenario #1
Execute the below command and pray it works well because you cannot enforce
\r\n due to max 1 char restriction.
{code:java}
from pyspark.sql.functions import *
df = (spark.read
.format("com.databricks.spark.csv")
.option("sep", "~")
.option("header", True)
.option("multiLine", True)
.option("inferSchema", False)
.option("encoding", 'utf-16')
.load("dbfs:/..../TestData.csv"))
{code}
Result : The result shows that the parsing didn't go well with multiLine option
!image-2021-08-26-14-04-47-464.png!
Scenario #2
Since \r\n string has \r character in it. Pass \r to the lineSep
{code:java}
from pyspark.sql.functions import *
df = (spark.read
.format("com.databricks.spark.csv")
.option("sep", "~")
.option("lineSep", '\r')
.option("header", True)
.option("multiLine", True)
.option("inferSchema", False)
.option("encoding", 'utf-16')
.load("dbfs:/FileStore/shared_uploads/[email protected]/TestData.csv"))
display(df)
{code}
Result : You get records parsed properly but the last character \n in the file
creates a blank line. This is the workaround for time being.
!image-2021-08-26-14-12-30-397.png!
> spark.read.csv is throwing exception ,"lineSep' can contain only 1 character"
> when parsing windows line feed (CR LF)
> --------------------------------------------------------------------------------------------------------------------
>
> Key: SPARK-34529
> URL: https://issues.apache.org/jira/browse/SPARK-34529
> Project: Spark
> Issue Type: Improvement
> Components: PySpark, SQL
> Affects Versions: 3.0.3, 3.1.1, 3.2.0
> Reporter: Shanmugavel Kuttiyandi Chandrakasu
> Priority: Minor
> Attachments: TestData.csv, image-2021-08-26-14-04-47-464.png,
> image-2021-08-26-14-06-41-055.png, image-2021-08-26-14-12-30-397.png
>
>
> lineSep documentation says -
> `lineSep` (default covers all `\r`, `\r\n` and `\n`): defines the line
> separator that should be used for parsing. Maximum length is 1 character.
> Reference:
>
> [https://spark.apache.org/docs/latest/api/python/pyspark.sql.html#pyspark.sql.DataFrameReader]
> When reading csv file using spark
> src_df = (spark.read
> .option("header", "true")
> .option("multiLine","true")
> .option("escape", "ǁ")
> .option("lineSep","\r\n")
> .schema(materialusetype_Schema)
> .option("badRecordsPath","/fh_badfile")
> .csv("<path-to-csv>/crlf.csv")
> )
> Below is the stack trace:
> java.lang.IllegalArgumentException: requirement failed: 'lineSep' can contain
> only 1 character.java.lang.IllegalArgumentException: requirement failed:
> 'lineSep' can contain only 1 character. at
> scala.Predef$.require(Predef.scala:281) at
> org.apache.spark.sql.catalyst.csv.CSVOptions.$anonfun$lineSeparator$1(CSVOptions.scala:209)
> at scala.Option.map(Option.scala:230) at
> org.apache.spark.sql.catalyst.csv.CSVOptions.<init>(CSVOptions.scala:207) at
> org.apache.spark.sql.catalyst.csv.CSVOptions.<init>(CSVOptions.scala:58) at
> org.apache.spark.sql.execution.datasources.csv.CSVFileFormat.buildReader(CSVFileFormat.scala:108)
> at
> org.apache.spark.sql.execution.datasources.FileFormat.buildReaderWithPartitionValues(FileFormat.scala:132)
> at
> org.apache.spark.sql.execution.datasources.FileFormat.buildReaderWithPartitionValues$(FileFormat.scala:123)
> at
> org.apache.spark.sql.execution.datasources.TextBasedFileFormat.buildReaderWithPartitionValues(FileFormat.scala:162)
> at
> org.apache.spark.sql.execution.FileSourceScanExec.inputRDD$lzycompute(DataSourceScanExec.scala:510)
> at
> org.apache.spark.sql.execution.FileSourceScanExec.inputRDD(DataSourceScanExec.scala:497)
> at
> org.apache.spark.sql.execution.FileSourceScanExec.doExecute(DataSourceScanExec.scala:692)
> at
> org.apache.spark.sql.execution.SparkPlan.$anonfun$execute$1(SparkPlan.scala:196)
> at
> org.apache.spark.sql.execution.SparkPlan.$anonfun$executeQuery$1(SparkPlan.scala:240)
> at
> org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:165)
> at
> org.apache.spark.sql.execution.SparkPlan.executeQuery(SparkPlan.scala:236) at
> org.apache.spark.sql.execution.SparkPlan.execute(SparkPlan.scala:192) at
> org.apache.spark.sql.execution.collect.Collector$.collect(Collector.scala:79)
> at
> org.apache.spark.sql.execution.collect.Collector$.collect(Collector.scala:88)
> at
> org.apache.spark.sql.execution.collect.InternalRowFormat$.collect(cachedSparkResults.scala:61)
> at
> org.apache.spark.sql.execution.collect.InternalRowFormat$.collect(cachedSparkResults.scala:57)
> at
> org.apache.spark.sql.execution.ResultCacheManager.$anonfun$getOrComputeResultInternal$1(ResultCacheManager.scala:483)
> at scala.Option.getOrElse(Option.scala:189) at
> org.apache.spark.sql.execution.ResultCacheManager.getOrComputeResultInternal(ResultCacheManager.scala:483)
> at
> org.apache.spark.sql.execution.ResultCacheManager.getOrComputeResult(ResultCacheManager.scala:427)
> at
> org.apache.spark.sql.execution.CollectLimitExec.executeCollectResult(limit.scala:58)
> at org.apache.spark.sql.Dataset.collectResult(Dataset.scala:3013) at
> org.apache.spark.sql.Dataset.$anonfun$collectResult$1(Dataset.scala:3004) at
> org.apache.spark.sql.Dataset.$anonfun$withAction$1(Dataset.scala:3728) at
> org.apache.spark.sql.execution.SQLExecution$.$anonfun$withCustomExecutionEnv$5(SQLExecution.scala:116)
> at
> org.apache.spark.sql.execution.SQLExecution$.withSQLConfPropagated(SQLExecution.scala:248)
> at
> org.apache.spark.sql.execution.SQLExecution$.$anonfun$withCustomExecutionEnv$1(SQLExecution.scala:101)
> at org.apache.spark.sql.SparkSession.withActive(SparkSession.scala:841) at
> org.apache.spark.sql.execution.SQLExecution$.withCustomExecutionEnv(SQLExecution.scala:77)
> at
> org.apache.spark.sql.execution.SQLExecution$.withNewExecutionId(SQLExecution.scala:198)
> at org.apache.spark.sql.Dataset.withAction(Dataset.scala:3726) at
> org.apache.spark.sql.Dataset.collectResult(Dataset.scala:3003)
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
