Package org.apache.hadoop.ozone.container.metadata

Class SchemaOneKeyCodec

java.lang.Object
org.apache.hadoop.ozone.container.metadata.SchemaOneKeyCodec
All Implemented Interfaces:
org.apache.hadoop.hdds.utils.db.Codec<String>
public final class SchemaOneKeyCodec extends Object implements org.apache.hadoop.hdds.utils.db.Codec<String>
Containers written using schema version 1 wrote unprefixed block ID keys as longs, and metadata or prefixed block IDs as Strings. This was done before codecs were introduced, so callers could serialize their data however they wanted. This codec handles this by checking for string prefixes in the data, and determining which format it should be encoded/decoded to/from.

  • Field Summary

    Fields inherited from interface org.apache.hadoop.hdds.utils.db.Codec

    EMPTY_BYTE_ARRAY

  • Method Summary

    Modifier and Type
    Method
    Description
    String
    copyObject(String object)
     
    String
    fromPersistedFormat(byte[] rawData)
    Determines whether the byte array was originally stored as a string or a long, decodes the data in the corresponding format, and returns it as a String.
    static org.apache.hadoop.hdds.utils.db.Codec<String>
    get()
     
    Class<String>
    getTypeClass()
     
    byte[]
    toPersistedFormat(String stringObject)
     

    Methods inherited from class java.lang.Object

    clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

    Methods inherited from interface org.apache.hadoop.hdds.utils.db.Codec

    fromCodecBuffer, fromPersistedFormatImpl, supportCodecBuffer, toCodecBuffer, toDirectCodecBuffer, toHeapCodecBuffer, toPersistedFormatImpl

  • Method Details

    • get

      public static org.apache.hadoop.hdds.utils.db.Codec<String> get()

    • getTypeClass

      public Class<String> getTypeClass()
      Specified by:
      getTypeClass in interface org.apache.hadoop.hdds.utils.db.Codec<String>

    • toPersistedFormat

      public byte[] toPersistedFormat(String stringObject) throws org.apache.hadoop.hdds.utils.db.CodecException
      Specified by:
      toPersistedFormat in interface org.apache.hadoop.hdds.utils.db.Codec<String>
      Throws:
      org.apache.hadoop.hdds.utils.db.CodecException

    • fromPersistedFormat

      public String fromPersistedFormat(byte[] rawData)
      Determines whether the byte array was originally stored as a string or a long, decodes the data in the corresponding format, and returns it as a String. Data is determined to have been stored as a string if it matches the known format of a metadata key or a prefixed block ID key. If the data does not match one of these formats, it will be parsed as a long only if it is 8 bytes long. Otherwise, it will be parsed as a String.

      Note that it is technically possible, although highly unlikely, that rawData was originally encoded as a long, but also happens to match the regex of a known string format when decoded. In this case this method will decode the data as a string. Log trace messages have been added to help debug these errors if necessary.

      Specified by:
      fromPersistedFormat in interface org.apache.hadoop.hdds.utils.db.Codec<String>
      Parameters:
      rawData - Byte array from the key/value store. Should not be null.

    • copyObject

      public String copyObject(String object)
      Specified by:
      copyObject in interface org.apache.hadoop.hdds.utils.db.Codec<String>