Muokkaa

Jaa


Troubleshoot the Parquet format connector in Azure Data Factory and Azure Synapse

APPLIES TO: Azure Data Factory Azure Synapse Analytics

Tip

Try out Data Factory in Microsoft Fabric, an all-in-one analytics solution for enterprises. Microsoft Fabric covers everything from data movement to data science, real-time analytics, business intelligence, and reporting. Learn how to start a new trial for free!

This article provides suggestions to troubleshoot common problems with the Parquet format connector in Azure Data Factory and Azure Synapse.

Error code: ParquetJavaInvocationException

  • Message: An error occurred when invoking java, message: %javaException;.

  • Causes and recommendations: Different causes may lead to this error. Check below list for possible cause analysis and related recommendation.

    Cause analysis Recommendation
    When the error message contains the strings "java.lang.OutOfMemory", "Java heap space", and "doubleCapacity", it's usually a memory management issue in an old version of Integration Runtime. If you are using Self-hosted IR and the version is earlier than 3.20.7159.1, we recommend that you upgrade to the latest version.
    When the error message contains the string "java.lang.OutOfMemory", the integration runtime doesn't have enough resources to process the files. Limit the concurrent runs on the integration runtime. For Self-hosted IR, scale up to a powerful machine with memory that's equal to or greater than 8 GB.
    When the error message contains the string "NullPointerReference", it might be a transient error. Retry the operation. If the problem persists, contact support.

Error code: ParquetInvalidFile

  • Message: File is not a valid Parquet file.

  • Cause: This is a Parquet file issue.

  • Recommendation: Check to see whether the input is a valid Parquet file.

Error code: ParquetNotSupportedType

Error code: ParquetMissedDecimalPrecisionScale

  • Message: Decimal Precision or Scale information is not found in schema for column: %column;.

  • Cause: The number precision and scale were parsed, but no such information was provided.

  • Recommendation: The source doesn't return the correct precision and scale information. Check the issue column for the information.

Error code: ParquetInvalidDecimalPrecisionScale

  • Message: Invalid Decimal Precision or Scale. Precision: %precision; Scale: %scale;.

  • Cause: The schema is invalid.

  • Recommendation: Check the issue column for precision and scale.

Error code: ParquetColumnNotFound

  • Message: Column %column; does not exist in Parquet file.

  • Cause: The source schema is a mismatch with the sink schema.

  • Recommendation: Check the mappings in the activity. Make sure that the source column can be mapped to the correct sink column.

Error code: ParquetInvalidDataFormat

  • Message: Incorrect format of %srcValue; for converting to %dstType;.

  • Cause: The data can't be converted into the type that's specified in mappings.source.

  • Recommendation: Double-check the source data or specify the correct data type for this column in the copy activity column mapping. For more information, see Supported file formats and compression codecs by the copy activity.

Error code: ParquetDataCountNotMatchColumnCount

  • Message: The data count in a row '%sourceColumnCount;' does not match the column count '%sinkColumnCount;' in given schema.

  • Cause: A mismatch between the source column count and the sink column count.

  • Recommendation: Double-check to ensure that the source column count is same as the sink column count in 'mapping'.

Error code: ParquetDataTypeNotMatchColumnType

  • Message: The data type %srcType; is not match given column type %dstType; at column '%columnIndex;'.

  • Cause: The data from the source can't be converted to the type that's defined in the sink.

  • Recommendation: Specify a correct type in mapping.sink.

Error code: ParquetBridgeInvalidData

  • Message: %message;

  • Cause: The data value has exceeded the limit.

  • Recommendation: Retry the operation. If the issue persists, contact us.

Error code: ParquetUnsupportedInterpretation

  • Message: The given interpretation '%interpretation;' of Parquet format is not supported.

  • Cause: This scenario isn't supported.

  • Recommendation: 'ParquetInterpretFor' should not be 'sparkSql'.

Error code: ParquetUnsupportFileLevelCompressionOption

  • Message: File level compression is not supported for Parquet.

  • Cause: This scenario isn't supported.

  • Recommendation: Remove 'CompressionType' in the payload.

Error code: UserErrorJniException

  • Message: Cannot create JVM: JNI return code [-6][JNI call failed: Invalid arguments.]

  • Cause: A Java Virtual Machine (JVM) can't be created because some illegal (global) arguments are set.

  • Recommendation: Log in to the machine that hosts each node of your self-hosted IR. Check to ensure that the system variable is set correctly, as follows: _JAVA_OPTIONS "-Xms256m -Xmx16g" with memory bigger than 8 G. Restart all the IR nodes, and then rerun the pipeline.

Arithmetic overflow

  • Symptoms: Error message occurred when you copy Parquet files: Message = Arithmetic Overflow., Source = Microsoft.DataTransfer.Common

  • Cause: Currently only the decimal of precision <= 38 and length of integer part <= 20 are supported when you copy files from Oracle to Parquet.

  • Resolution: As a workaround, you can convert any columns with this problem into VARCHAR2.

No enum constant

  • Symptoms: Error message occurred when you copy data to Parquet format: java.lang.IllegalArgumentException:field ended by &apos;;&apos;, or: java.lang.IllegalArgumentException:No enum constant org.apache.parquet.schema.OriginalType.test.

  • Cause:

    The issue could be caused by white spaces or unsupported special characters (such as,;{}()\n\t=) in the column name, because Parquet doesn't support such a format.

    For example, a column name such as contoso(test) will parse the type in brackets from code Tokenizer st = new Tokenizer(schemaString, " ;{}()\n\t");. The error is thrown because there is no such "test" type.

    To check supported types, go to the GitHub apache/parquet-mr site.

  • Resolution:

    Double-check to see whether:

    • There are white spaces in the sink column name.
    • The first row with white spaces is used as the column name.
    • The type OriginalType is supported. Try to avoid using these special characters: ,;{}()\n\t=.

Error code: ParquetDateTimeExceedLimit

  • Message: The Ticks value '%ticks;' for the datetime column must be between valid datetime ticks range -621355968000000000 and 2534022144000000000.

  • Cause: If the datetime value is '0001-01-01 00:00:00', it could be caused by the difference between Julian Calendar and Gregorian Calendar. For more details, reference Difference between Julian and proleptic Gregorian calendar dates.

  • Resolution: Check the ticks value and avoid using the datetime value '0001-01-01 00:00:00'.

Error code: ParquetInvalidColumnName

  • Message: The column name is invalid. Column name cannot contain these character:[,;{}()\n\t=]

  • Cause: The column name contains invalid characters.

  • Resolution: Add or modify the column mapping to make the sink column name valid.

The file created by the copy data activity extracts a table that contains a varbinary (max) column

  • Symptoms: The Parquet file created by the copy data activity extracts a table that contains a varbinary (max) column.

  • Cause: This issue is caused by the Parquet-mr library bug of reading large column.

  • Resolution: Try to generate smaller files (size < 1G) with a limitation of 1000 rows per file.

For more troubleshooting help, try these resources: