You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@iceberg.apache.org by GitBox <gi...@apache.org> on 2020/11/23 19:10:43 UTC
[GitHub] [iceberg] rdblue commented on a change in pull request #1802: Add delta writer classes
rdblue commented on a change in pull request #1802:
URL: https://github.com/apache/iceberg/pull/1802#discussion_r528934829
##########
File path: spark/src/main/java/org/apache/iceberg/spark/source/SparkAppenderFactory.java
##########
@@ -86,4 +98,81 @@
throw new RuntimeIOException(e);
}
}
+
+ @Override
+ public DataWriter<InternalRow> newDataWriter(EncryptedOutputFile file, FileFormat fileFormat, StructLike partition) {
+ return new DataWriter<>(
+ newAppender(file.encryptingOutputFile(), fileFormat), fileFormat,
+ file.encryptingOutputFile().location(), spec, partition, file.keyMetadata());
+ }
+
+ @Override
+ public EqualityDeleteWriter<InternalRow> newEqDeleteWriter(EncryptedOutputFile outputFile, FileFormat format,
+ StructLike partition) {
+ try {
+ switch (format) {
+ case PARQUET:
+ return Parquet.writeDeletes(outputFile.encryptingOutputFile())
+ .createWriterFunc(msgType -> SparkParquetWriters.buildWriter(dsSchema, msgType))
+ .overwrite()
+ .rowSchema(writeSchema)
+ .withSpec(spec)
+ .withPartition(partition)
+ .equalityFieldIds(equalityFieldIds)
+ .withKeyMetadata(outputFile.keyMetadata())
+ .buildEqualityWriter();
+
+ case AVRO:
+ return Avro.writeDeletes(outputFile.encryptingOutputFile())
+ .createWriterFunc(ignored -> new SparkAvroWriter(dsSchema))
+ .overwrite()
+ .rowSchema(writeSchema)
+ .withSpec(spec)
+ .withPartition(partition)
+ .equalityFieldIds(equalityFieldIds)
+ .withKeyMetadata(outputFile.keyMetadata())
+ .buildEqualityWriter();
+
+ default:
+ throw new UnsupportedOperationException("Cannot write unsupported format: " + format);
+ }
+
+ } catch (IOException e) {
+ throw new UncheckedIOException("Failed to create new equality delete writer", e);
+ }
+ }
+
+ @Override
+ public PositionDeleteWriter<InternalRow> newPosDeleteWriter(EncryptedOutputFile outputFile, FileFormat format,
+ StructLike partition) {
+ try {
+ switch (format) {
+ case PARQUET:
+ return Parquet.writeDeletes(outputFile.encryptingOutputFile())
+ .createWriterFunc(msgType -> SparkParquetWriters.buildWriter(dsSchema, msgType))
Review comment:
The builder here will automatically wrap the row schema and the function passed here to add the extra schema layer. So we just need to configure this for rows, not for the combined schema. That's part of why the position writer's delete method accepts file and position independent of row, to keep the encapsulation and not leak the concern to places like this.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org