Skip to content

Support for UUID columns #9

@jze

Description

@jze

I try to write a UUID column.

My first attempt to define the type was

types.add(Types.required(BINARY).as(LogicalTypeAnnotation.uuidType()).named("id"));

However, that does not work:

Exception in thread "main" java.lang.IllegalStateException: UUID can only annotate FIXED_LEN_BYTE_ARRAY(16)
	at org.apache.parquet.Preconditions.checkState(Preconditions.java:166)
	at org.apache.parquet.schema.Types$BasePrimitiveBuilder$1.checkFixedPrimitiveType(Types.java:594)
	at org.apache.parquet.schema.Types$BasePrimitiveBuilder$1.visit(Types.java:464)
	at org.apache.parquet.schema.LogicalTypeAnnotation$UUIDLogicalTypeAnnotation.accept(LogicalTypeAnnotation.java:1014)
	at org.apache.parquet.schema.Types$BasePrimitiveBuilder.build(Types.java:442)
	at org.apache.parquet.schema.Types$BasePrimitiveBuilder.build(Types.java:334)
	at org.apache.parquet.schema.Types$Builder.named(Types.java:313)

Next I tried to define the type like this:

types.add(Types.required(FIXED_LEN_BYTE_ARRAY).length(16).as(LogicalTypeAnnotation.uuidType()).named("id"));

The definition of the schema works. However, when I try to write a value

byte[] uuidBytes = ByteBuffer.allocate(16)
                    .putLong(uuid.getMostSignificantBits())
                    .putLong(uuid.getLeastSignificantBits())
                    .array();
valueWriter.write("id", uuidBytes);

it fails because FIXED_LEN_BYTE_ARRAY is not supported:

Exception in thread "main" java.lang.UnsupportedOperationException: We don't support writing FIXED_LEN_BYTE_ARRAY
	at blue.strategic.parquet.ParquetWriter$SimpleWriteSupport.writeField(ParquetWriter.java:161)

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions