ghkang98 commented on code in PR #36289:
URL: https://github.com/apache/doris/pull/36289#discussion_r1642061231
##########
fe/fe-core/src/main/java/org/apache/doris/datasource/iceberg/IcebergTransaction.java:
##########
@@ -59,140 +66,114 @@ public void
updateIcebergCommitData(List<TIcebergCommitData> commitDataList) {
}
}
- public void beginInsert(String dbName, String tbName) {
- Table icebergTable =
ops.getCatalog().loadTable(TableIdentifier.of(dbName, tbName));
- transaction = icebergTable.newTransaction();
+ public void pendingCommit(SimpleTableInfo tableInfo) {
+ this.tableInfo = tableInfo;
+ this.transaction = getNativeTable(tableInfo).newTransaction();
}
- public void finishInsert() {
- Table icebergTable = transaction.table();
- AppendFiles appendFiles = transaction.newAppend();
-
- for (CommitTaskData task : convertToCommitTaskData()) {
- DataFiles.Builder builder = DataFiles.builder(icebergTable.spec())
- .withPath(task.getPath())
- .withFileSizeInBytes(task.getFileSizeInBytes())
- .withFormat(IcebergUtils.getFileFormat(icebergTable))
- .withMetrics(task.getMetrics());
-
- if (icebergTable.spec().isPartitioned()) {
- List<String> partitionValues = task.getPartitionValues()
- .orElseThrow(() -> new VerifyException("No partition
data for partitioned table"));
- builder.withPartitionValues(partitionValues);
- }
- appendFiles.appendFile(builder.build());
- }
+ public void preCommit(SimpleTableInfo tableInfo,
Optional<InsertCommandContext> insertCtx) {
- // in appendFiles.commit, it will generate metadata(manifest and
snapshot)
- // after appendFiles.commit, in current transaction, you can already
see the new snapshot
- appendFiles.commit();
- }
+ LOG.info("iceberg table {} insert table finished!", tableInfo);
- public List<CommitTaskData> convertToCommitTaskData() {
- List<CommitTaskData> commitTaskData = new ArrayList<>();
- for (TIcebergCommitData data : this.commitDataList) {
- commitTaskData.add(new CommitTaskData(
- data.getFilePath(),
- data.getFileSize(),
- new Metrics(
- data.getRowCount(),
- Collections.EMPTY_MAP,
- Collections.EMPTY_MAP,
- Collections.EMPTY_MAP,
- Collections.EMPTY_MAP
- ),
- data.isSetPartitionValues() ?
Optional.of(data.getPartitionValues()) : Optional.empty(),
- convertToFileContent(data.getFileContent()),
- data.isSetReferencedDataFiles() ?
Optional.of(data.getReferencedDataFiles()) : Optional.empty()
- ));
+ //create start the iceberg transaction
+ TUpdateMode updateMode = TUpdateMode.APPEND;
+ if (insertCtx.isPresent()) {
+ updateMode = ((BaseExternalTableInsertCommandContext)
insertCtx.get()).isOverwrite() ? TUpdateMode.OVERWRITE
+ : TUpdateMode.APPEND;
}
- return commitTaskData;
+ updateManifestAfterInsert(updateMode);
}
- private FileContent convertToFileContent(TFileContent content) {
- if (content.equals(TFileContent.DATA)) {
- return FileContent.DATA;
- } else if (content.equals(TFileContent.POSITION_DELETES)) {
- return FileContent.POSITION_DELETES;
+ private void updateManifestAfterInsert(TUpdateMode updateMode) {
+
+ Table table = getNativeTable(tableInfo);
Review Comment:
First, the getNativeTable function gets the table from IcebergUtil.
IcebergUtil is a tool class that is not only used in the current scenario.
Secondly, we should also need to get the latest information of the table.
--
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.
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
